All notable changes to this project will be documented in this file. See commit-and-tag-version for commit guidelines.
- [Private Aggregation] Make ig index available to populate interest_group_index in PrivateAggrgeteContribution
- [Private Aggregation] Populate id_idx for PrivateAggregateContribution
- Add allow_compressed_auction_config flag to Terraform
- Add bidding.business_logic.debug_url_count metric
- Add bidding.business_logic.debug_urls_size_bytes metric
- Add bidding.business_logic.failed_to_bid_percent metric
- Add compressed auction config field to API
- Add dimension for GenerateBids timed out errors to bidding.errors_count
- Add TCMalloc config for B&A inference
- Allow model eviction grace period
- Control whether scoring signals are required for scoring via flag
- Enable TCMalloc for the TensorFlow sidecar
- Read compressed AuctionConfig in SFE
- remove metric noise for sampled debug request
- Report model eviction via metrics
- With Flag Allow Just Component Signals
- Allocate compression output buffer on heap
- Missing inference metric label in AWS dashboard
- Set default value for PRESUBMIT_PERF to 0
- Standardize business logic metrics for generateBids PA and PAS
- Update tag to correctly pass roma_timeout_ms value to Roma
- [BYOB] Add guide and sample request for generateBid SDK
- Add consented debug logs to seller KV V2
- Add Data Version for buyer to generateBid() browser signals
- add egress schema bucket fetcher
- add egress schema fetch manager
- add flag https_fetch_skips_tls_verification for local testing
- Add priority fields to B&A API
- Add PV support on BFE
- Add PV support on SFE
- Add Seller Data Version to ReportResult
- Add Seller DV to Bid Metadata for ScoreAd()
- Add support for DNS routing by regional LB latency in AWS
- Add support for DNS routing by regional LB latency in AWS
- add top level seller debugging data
- Data Version in BuyerReporting
- enable seller debugging for eligible request
- Implement delete model rpc for inference sidecar
- Implement KV V2 Adapter for Scoring Signals
- Implement KV V2 flag for seller
- Implement v2 bidding support - config flags propagation
- improve SFE response debug logging
- Launch the B&A inference for both PA and PAS
- Pass Data Version to AuctionService
- Pass DV Hdr to Auction Service
- Runtime flag to make all non-prod requests as always consented
- Set Data Version on AdWithBid
- Support Data Version for Seller BYOS KV
- Support inference model deletion in the model fetcher
- support per-request spec of egress schema
- support per-request specification of seller and buyer udf versions using bucket blobs
- update proto definition for multi blob support
- Use separate tf parameter/endpoint for buyer TKV
- Use separate tf parameter/endpoint for seller TKV
- Validate DV Hdr val size based on clientType
- Validate DV Header values for size based on client type
- Check for key in perBuyerSignal before looking up buyer_signals for reporting
- Fix failing e2e test
- Fix Terraform apply when extra args are empty
- Load reportWin code when fetch_mode is local
- load the wrapper with code isolation for local fetch mode
- Log inference model registration failure reason
- Replace prefix-based matching with exact matching for model paths
- Revert saving and deleting Tensorflow models in place to fix memory leak
- Set the correct version_string for reportResult when udf loaded from bucket
- Support a wildcar SSL certificate in e2e benchmark
- update GCP cloudbuild with correct git tag name
- Use correct timeout for Kv V2 bidding signals call
- Add priority fields to B&A API
- Add support for DNS routing by regional LB latency in AWS
- Data Version in BuyerReporting
- Implement delete model rpc for inference sidecar
- Implement v2 bidding support - config flags propagation
- Launch the B&A inference for both PA and PAS
- Pass Data Version to AuctionService
- Set Data Version on AdWithBid
- Validate DV Header values for size based on client type
- Log inference model registration failure reason
- update GCP cloudbuild with correct git tag name
- [BYOB] Add udf_execution metrics for generateBid
- add proto field to enable buyer debugging for eligible request
- adding aws inference dashboard
- Enable B&A experiments in GCP by splitting traffic
- enable buyer debugging for eligible request
- Implement Data Version Header for KV Server Clients
- Implement v2 bidding support - add branch selection
- Implement v2 bidding support - add kv_buyer_signals_adapter
- Implement v2 bidding support - update kv dependency
- output packer logs to stderr on AMI creation failure
- Pass data version to GenerateBidsRawRequest
- Support model deletion in model store
- Update Kv client to match the latest contract
- Accumulate js_execution.errors_count metric
- Copy common signals instead of move in BuildProtectedAudienceBidRequest
- Fix changelog generation
- GCP demo terraform outputs url
- Prevent a potential read-after-delete bug in the inference benchmark
- start_bidding runs inference default js file
- typo in bidding_service_integration_test and buyer_code_wrapper
- [BYOB] Add buyer code fetcher for executables
- [BYOB] Add dispatch client for generateBid
- [BYOB] Add GCP build setup for Bidding service
- [BYOB] Add gRPC reactor for generateBid binary
- [BYOB] Add info about runtime flags to demo/README.md
- [BYOB] Add interface for clients that load code
- [BYOB] Add to Bidding service critical path
- Add a generic model validator interface for Tensorflow and PyTorch
- Add PrivateAggregation protos to generateBid spec
- add project setup scripts
- add proto support for updateIfOlderThanMs
- BFE parses perInterestGroupData, returns updateIfOlderThanMs data to SFE
- Compress SFE <> BFE comms
- Create new inference dashboard for GCP
- Enable chaffing on SFE prod mode builds
- Enable connecting to TKV via an External LB
- Enable dynamic partition for inference metrics
- Enable model validation for PyTorch
- Enable TCMalloc for B&A inference
- Export all available models as metric
- Freeze a TensorFlow model (bcc798c), closes #7
- Implement a PyTorch model validator
- Implement a TensorFlow model validator
- Rename JS_NUM_WORKERS to UDF_NUM_WORKERS
- SFE sends updateIfOlderThanMs data to client
- update start_bidding with minimal defaults and local file flag
- BYOB generateBid should only parse first bid from binary response
- Ensure gRPC calls don't hang on error
- Fix js_execution.error.count metric
- Fix memory leak in BFE client
- Fix readonly declarations in shell script
- Fix version bump in release script
- Load request files in the init function
- Reduce absl::GetFlag() call in GetBidsReactor
- Renames prev_wins_ms to prev_wins
- Set inference sidecar GRPC keepalive timeout
- update tls cert commands for GCP setup
- Use std::hash instead of absl::hash
- Move serialization of bidding signals per IG from Bidding to BFE
- [IsolateBuyerAndSellerCodeExecution] Add wrapper for buyer's reportWin() udf
- [PAS threat mitigation] Add support for egressVector and temporaryUnlimittedEgressVector by replacing the $extraArgs
- [Private Aggregation] Add all the contributions from Auction Service to AuctionResult
- [Private Aggregation] Add integration test to verify parsing of PAAPI event
- [Private Aggregation] Add PrivateAggregateReportingResponse to AuctionResult proto
- [Private Aggregation] Convert 128 bit bucket into array of 64 bit integer in wrapper
- [Private Aggregation] Convert bucket offset object to SignalBucket
- [Private Aggregation] Convert value object in contribution to SignalValue
- [PrivateAggregation] Add PrivateAggregateReportingResponse in Auction Service response
- [PrivateAggregation] Append private aggregation wrapper to existing seller's wrapper
- [reporting] Enable reportResult execution for topLevelSeller
- Add
common.privateAggregation.createContribution
function for creating a contribution object. - Add
common.privateAggregation.ReservedEvent
object that holds reserved events string constants. - Add a script to freeze a given TensorFlow model
- Add a script to list all TensorFlow ops
- Add a script to list all Torchscript ops
- Add AppendAdEventContributionsToPaggResponse with SignalValue only
- Add auction_service.privateAggregation.contributeToHistogram/OnEvent
- Add bucket protos (Bucket128Bit, BucketOffset, SignalBucket, PrivateAggregationBucket) and add bucket and value fields to PrivateAggregateContribution
- Add chaffing feature flags in SFE/BFE.
- Add checksum capability to model loading
- Add common utils to create test private aggregation contribution and response
- Add config flag enable_private_aggregate_reporting to bidding service's runtime_config
- Add config flag enable_private_aggregation_generation to auction service's runtime_config
- Add ConvertSignalValueToInt for calculating final value from contribution's base value, scale, and offset.
- Add distribution of IGs related metrics
- Add enable_private_aggregate_reporting for GetSellerWrappedCode and unit tests that uses GetSellerWrappedCode
- Add event code and private aggregation objects for contributeToHistogramOnEvent, also added tests for event field
- add EventMessage to log context
- Add HandlePrivateAggregateReporting with support of SignalValue only
- Add HandlePrivateAggregationContributions, iterate over AdWithBids, filter contributions, post processing, test
- Add helper functions convertEventToInt and convertBaseValueToInt and tests for them for contributeToHistogramOnEvent for bidding service
- Add Inference Metrics like count, duration, error and size to TF & Pytorch sidecars.
- Add Inference Request count, duration and failure metric partitioned by model path for Tensorflow and Pytorch sidecars.
- Add metrics for inference model fetching and registration
- add option to switch DebugInfo in secure invoke encrypt
- Add parser function for Private Aggregation's Signal Bucket Object
- Add parser function for Private Aggregation's Signal Value Object
- Add PrivateAggregateContribution proto with event field and enum EventType to the .proto file
- Add proto messages for SignalValue, PrivateAggregationValue, and BaseValue enum.
- Add request creation timestamp to ProtectedAuctionInput
- Add request/response to ExecuteInternal on async clients
- Add support for collecting metrics from the Roma callbacks.
- Add tee-container-log-redirect option in terraform
- Add ToBaseValue Helper Function to convert corresponding base value strings to BaseValue Enum
- Add ToEventTypeString for converting EventType enum into its corresponding string in JSON object returned by ROMA
- Add utility function to calculate Signal Bucket's final value post auction (3da9ab5), closes 1#L192
- Add utility function to parse and return BucketOffset from rapidjson document
- Add wrapper and test files with method headers for contributeToHistogramOnEvent for bidding service
- Added
isValidCustomEvent
validation function for common.privateAggregation - Added isValidValue validation function for common.privateAggregation
- Added new files for JS private aggregation util for bidding service
- Bash script for ASG and Cloud Map Custom HealthChecks
- Bid Currency Support for Top-Level Seller in Server-Orchestrated Multi-Seller Auctions
- Change js helper convertEventToInt to mapEventToEnum to return string corresponding to enum instead of int
- configured public key urls are verified against an explicit allowlist
- Consented request replace enable_adtech_code_logging in Bidding Server
- contributeToHistogram with test for bidding service
- contributeToHistogramOnEvent with tests, using common private aggregation helpers
- convert all remaining docker images into OCI
- Convert Private Aggregation wrapper functions from .js to C++ string
- Create a periodic model fetcher library for inference
- Create declarative spec for generateBid
- Create inference model store for model management
- Create logging library for inference consented logs
- Debug Reporting for Bid Currency
- Declare inference JS error schema
- Declare schema for model fetching metadata file
- decrease aws/build_and_test duration by ~75%
- Enable blob fetcher to only fetch blobs with given prefixes
- Enable dynamic model loading for inference
- Enable Protected App Signals by Default
- Enable Service Mesh on AWS by Default
- export non-privacy log through otel with safe system context
- Feature Flag for TLS in Service Mesh
- Force chaffing to enabled for prod builds on BFE + update common repo dep
- Force the ML model reset with the probability of 0.1%
- Forward per request consented debugging config to roma callback
- Freeze TorchScript models before serving
- Handle chaff and new request format on BFE
- Implement aggregated Error Reporting for Tensorflow and Pytorch sidecars.
- Implement AWS Cloud Un-Map
- Implement new SFE <> BFE request format for chaffing
- Implement ParseAndProcessContribution with SignalValue only
- Implement the probabilistic model reset for PyTorch
- Implement the probabilistic model reset for TensorFlow.
- Instrument inference JS error for the PyTorch sidecar
- Integrate with Trusted KV Server in Mesh
- Load models using model metadata config for the periodic model fetcher
- Load Test Flag for AWS
- Log consented debugging information in inference sidecar
- log EventMessage in servers
- log udf log in EventMessage for non_prod debug_info
- Make chaffing flag configurable via Terraform
- move ig metric to bfe
- Move serialization of bidding signals per IG from Bidding to BFE
- Propagate inference error back to JS caller
- Remove enableAdtechCodeLogging flag value
- Remove Envoy Access Logging
- Route consented inference requests to a consented model store
- send chaff requests from SFE
- Service Mesh in AWS
- Support CPU isolation in the inference sidecar
- Unit test for loading parsing fake contributions in Roma for bidding service
- Unpad KV server responses
- update code/cloud build to use tags
- Update Demo Terraform Configs with values for integrating TEE KV Servers into Mesh
- Upgrade AWS Provider for Terraform from v3.xx to v4.xx
- Use gRPC for AWS Service Mesh Envoy HCs
- Write a fake generateBid() script which calls privateAggregation.contributeToHistogramOnEvent(event, )
- write event message to gcs for consented request
- write event message to s3 in aws for consented request
- [IsolateBuyerAndSellerCodeExecution] Add a new class for buyer's reporting code fetch and load
- [IsolateBuyerAndSellerCodeExecution] Add a new code wrapper with only scoreAd and reportResult
- [IsolateBuyerAndSellerCodeExecution] Add config flag to enable seller and buyer code isolation
- [IsolateBuyerAndSellerCodeExecution] Modify seller_udf_manager to fetch and load buyer udfs
- [IsolateBuyerAndSellerCodeExecution] Refactor the code fetch files.
- Add BidCurrency SUT to Pre-Submit
- add check back in default grpc client
- Add createContribution inside contributeToHistogram/OnEvent and fixed typo with createContribution's value conditional statement.
- Add DebugInfo pointer for debugging log into RomaRequestContext
- Add generation_id to chaff requests
- Add handling when rejection reason is specified as base value but not available and statuscode documentation for HandlePrivateAggregationValuePostAuction
- Adds domain equality validations to buyer reporting UDF URL
- Align plaintext buyer request for basic SUT with encrypted version
- Allow up to 15 buyers when chaffing enabled
- chaffing bugs (027d5bb), closes /github.com/abseil/abseil-cpp/blob/master/absl/container/internal/raw_hash_set.h#L1572
- Changed base value string to be converted in ToBaseValue function and test cases
- Changed raw string delimiter JSCODE to JS_CODE for consistency
- Changed the input format of HandlePrivateAggregationReporting, added support for parsing PAgg bucket, and add the logic for getting required BaseValues (winning-bid, highest-scoring-other-bid, rejection-reason).
- clean up log verbosity 3
- Correct output_filter typo
- do not try to impersonate service accounts if TEST_MODE=true
- Don't set chaff size for non-chaff requests
- Eliminate Terraform Error Message about empty Authority Field
- Enable threat mitigation with seller and buyer code isolation
- Ensure instance id is set in logs on AWS when not using mesh.
- Execute Callback for empty HTTP request vector
- Fix bugs in reportWin URL validation in auction service
- Fix release notes by adding a dedicated tag on main to generate the changelog
- make num_chaff_requests not have a static lower bound
- Make terraform for aws delete cloud maps without error
- Populate temp rc file for inference sidecar
- Populate temp rc file for inference sidecar
- Populate temp rc file for inference sidecar
- Redirect misleading log to /dev/null
- Refactored conversion of private aggregation value post auction into returning absl::StatusOr instead of implicitly editing PrivateAggregationValue that was passed in.
- Revert wrk2 test runner to an earlier version
- scorecard.yaml version updates
- shuffle request order (real and chaff) on SFE
- truncate curl failure info for fail count metric
- undo rules_oci migration to fix hash stability issues
- Update setup_2 and demo terraform configs to valid state
- Update SUTs to include and test Experiment Group ID
- Upgrades google terraform plugin to 5.31.0 to fix crash
- Add PAS input example to BFE
- Document how to deployment B&A without inference
- Update inference_sidecar README with model size limits
- Updated expected format for INFERENCE_MODEL_BUCKET_PATHS in README
- [IsolateBuyerAndSellerCodeExecution] Add wrapper for buyer's reportWin() udf
- [PrivateAggregation] Add PrivateAggregateReportingResponse in Auction Service response
- [PrivateAggregation] Append private aggregation wrapper to existing seller's wrapper
- [reporting] Enable reportResult execution for topLevelSeller
- Add
common.privateAggregation.createContribution
function for creating a contribution object. - Add
common.privateAggregation.ReservedEvent
object that holds reserved events string constants. - Add auction_service.privateAggregation.contributeToHistogram/OnEvent
- Add bucket protos (Bucket128Bit, BucketOffset, SignalBucket, PrivateAggregationBucket) and add bucket and value fields to PrivateAggregateContribution
- Add chaffing feature flags in SFE/BFE.
- Add checksum capability to model loading
- Add config flag enable_private_aggregate_reporting to bidding service's runtime_config
- Add config flag enable_private_aggregation_generation to auction service's runtime_config
- Add ConvertSignalValueToInt for calculating final value from contribution's base value, scale, and offset.
- Add distribution of IGs related metrics
- Add enable_private_aggregate_reporting for GetSellerWrappedCode and unit tests that uses GetSellerWrappedCode
- Add event code and private aggregation objects for contributeToHistogramOnEvent, also added tests for event field
- add EventMessage to log context
- Add helper functions convertEventToInt and convertBaseValueToInt and tests for them for contributeToHistogramOnEvent for bidding service
- Add Inference Metrics like count, duration, error and size to TF & Pytorch sidecars.
- add option to switch DebugInfo in secure invoke encrypt
- Add PrivateAggregateContribution proto with event field and enum EventType to the .proto file
- Add proto messages for SignalValue, PrivateAggregationValue, and BaseValue enum.
- Add request creation timestamp to ProtectedAuctionInput
- Add request/response to ExecuteInternal on async clients
- Add support for collecting metrics from the Roma callbacks.
- Add tee-container-log-redirect option in terraform
- Add wrapper and test files with method headers for contributeToHistogramOnEvent for bidding service
- Added
isValidCustomEvent
validation function for common.privateAggregation - Added isValidValue validation function for common.privateAggregation
- Added new files for JS private aggregation util for bidding service
- Bash script for ASG and Cloud Map Custom HealthChecks
- Bid Currency Support for Top-Level Seller in Server-Orchestrated Multi-Seller Auctions
- Change js helper convertEventToInt to mapEventToEnum to return string corresponding to enum instead of int
- configured public key urls are verified against an explicit allowlist
- Consented request replace enable_adtech_code_logging in Bidding Server
- contributeToHistogram with test for bidding service
- contributeToHistogramOnEvent with tests, using common private aggregation helpers
- Convert Private Aggregation wrapper functions from .js to C++ string
- Create a periodic model fetcher library for inference
- Create inference model store for model management
- Create logging library for inference consented logs
- Debug Reporting for Bid Currency
- Declare inference JS error schema
- Declare schema for model fetching metadata file
- decrease aws/build_and_test duration by ~75%
- Enable blob fetcher to only fetch blobs with given prefixes
- Enable dynamic model loading for inference
- Enable Service Mesh on AWS by Default
- export non-privacy log through otel with safe system context
- Feature Flag for TLS in Service Mesh
- Force chaffing to enabled for prod builds on BFE + update common repo dep
- Force the ML model reset with the probability of 0.1%
- Forward per request consented debugging config to roma callback
- Handle chaff and new request format on BFE
- Implement AWS Cloud Un-Map
- Implement new SFE <> BFE request format for chaffing
- Implement the probabilistic model reset for PyTorch
- Implement the probabilistic model reset for TensorFlow.
- Instrument inference JS error for the PyTorch sidecar
- Integrate with Trusted KV Server in Mesh
- Load models using model metadata config for the periodic model fetcher
- Load Test Flag for AWS
- Log consented debugging information in inference sidecar
- log EventMessage in servers
- log udf log in EventMessage for non_prod debug_info
- Make chaffing flag configurable via Terraform
- move ig metric to bfe
- Propagate inference error back to JS caller
- Remove enableAdtechCodeLogging flag value
- Remove Envoy Access Logging
- Route consented inference requests to a consented model store
- send chaff requests from SFE
- Service Mesh in AWS
- Support CPU isolation in the inference sidecar
- Unit test for loading parsing fake contributions in Roma for bidding service
- Unpad KV server responses
- update code/cloud build to use tags
- Update Demo Terraform Configs with values for integrating TEE KV Servers into Mesh
- Upgrade AWS Provider for Terraform from v3.xx to v4.xx
- Use gRPC for AWS Service Mesh Envoy HCs
- Write a fake generateBid() script which calls privateAggregation.contributeToHistogramOnEvent(event, )
- [IsolateBuyerAndSellerCodeExecution] Add a new class for buyer's reporting code fetch and load
- [IsolateBuyerAndSellerCodeExecution] Add a new code wrapper with only scoreAd and reportResult
- [IsolateBuyerAndSellerCodeExecution] Add config flag to enable seller and buyer code isolation
- [IsolateBuyerAndSellerCodeExecution] Modify seller_udf_manager to fetch and load buyer udfs
- [IsolateBuyerAndSellerCodeExecution] Refactor the code fetch files.
- Add BidCurrency SUT to Pre-Submit
- add check back in default grpc client
- Add createContribution inside contributeToHistogram/OnEvent and fixed typo with createContribution's value conditional statement.
- Add DebugInfo pointer for debugging log into RomaRequestContext
- Add generation_id to chaff requests
- Adds domain equality validations to buyer reporting UDF URL
- Align plaintext buyer request for basic SUT with encrypted version
- Allow up to 15 buyers when chaffing enabled
- chaffing bugs (027d5bb), closes /github.com/abseil/abseil-cpp/blob/master/absl/container/internal/raw_hash_set.h#L1572
- Changed raw string delimiter JSCODE to JS_CODE for consistency
- clean up log verbosity 3
- Correct output_filter typo
- do not try to impersonate service accounts if TEST_MODE=true
- Don't set chaff size for non-chaff requests
- Eliminate Terraform Error Message about empty Authority Field
- Ensure instance id is set in logs on AWS when not using mesh.
- Execute Callback for empty HTTP request vector
- Fix bugs in reportWin URL validation in auction service
- Fix release notes by adding a dedicated tag on main to generate the changelog
- make num_chaff_requests not have a static lower bound
- Make terraform for aws delete cloud maps without error
- Redirect misleading log to /dev/null
- scorecard.yaml version updates
- shuffle request order (real and chaff) on SFE
- truncate curl failure info for fail count metric
- Update setup_2 and demo terraform configs to valid state
- Update SUTs to include and test Experiment Group ID
- Upgrades google terraform plugin to 5.31.0 to fix crash
- Add PAS input example to BFE
- Document how to deployment B&A without inference
- Update inference_sidecar README with model size limits
- Updated expected format for INFERENCE_MODEL_BUCKET_PATHS in README
- [IsolateBuyerAndSellerCodeExecution] Add wrapper for buyer's reportWin() udf
- [reporting] Enable reportResult execution for topLevelSeller
- Add chaffing feature flags in SFE/BFE.
- Add config flag enable_private_aggregation_generation to auction service's runtime_config
- Add distribution of IGs related metrics
- add EventMessage to log context
- Add request creation timestamp to ProtectedAuctionInput
- Add request/response to ExecuteInternal on async clients
- Add support for collecting metrics from the Roma callbacks.
- Add tee-container-log-redirect option in terraform
- Bash script for ASG and Cloud Map Custom HealthChecks
- Bid Currency Support for Top-Level Seller in Server-Orchestrated Multi-Seller Auctions
- configured public key urls are verified against an explicit allowlist
- Consented request replace enable_adtech_code_logging in Bidding Server
- Create inference model store for model management
- Create logging library for inference consented logs
- Debug Reporting for Bid Currency
- decrease aws/build_and_test duration by ~75%
- Feature Flag for TLS in Service Mesh
- Force the ML model reset with the probability of 0.1%
- Forward per request consented debugging config to roma callback
- Handle chaff and new request format on BFE
- Implement AWS Cloud Un-Map
- Implement new SFE <> BFE request format for chaffing
- Implement the probabilistic model reset for PyTorch
- Implement the probabilistic model reset for TensorFlow.
- Integrate with Trusted KV Server in Mesh
- Load Test Flag for AWS
- Log consented debugging information in inference sidecar
- log EventMessage in servers
- log udf log in EventMessage for non_prod debug_info
- Remove enableAdtechCodeLogging flag value
- Remove Envoy Access Logging
- Route consented inference requests to a consented model store
- send chaff requests from SFE
- Service Mesh in AWS
- Support CPU isolation in the inference sidecar
- Unpad KV server responses
- update code/cloud build to use tags
- Upgrade AWS Provider for Terraform from v3.xx to v4.xx
- Use gRPC for AWS Service Mesh Envoy HCs
- [IsolateBuyerAndSellerCodeExecution] Add a new class for buyer's reporting code fetch and load
- [IsolateBuyerAndSellerCodeExecution] Add a new code wrapper with only scoreAd and reportResult
- [IsolateBuyerAndSellerCodeExecution] Add config flag to enable seller and buyer code isolation
- [IsolateBuyerAndSellerCodeExecution] Modify seller_udf_manager to fetch and load buyer udfs
- [IsolateBuyerAndSellerCodeExecution] Refactor the code fetch files.
- Add BidCurrency SUT to Pre-Submit
- add check back in default grpc client
- Add DebugInfo pointer for debugging log into RomaRequestContext
- Add generation_id to chaff requests
- Adds domain equality validations to buyer reporting UDF URL
- Align plaintext buyer request for basic SUT with encrypted version
- clean up log verbosity 3
- do not try to impersonate service accounts if TEST_MODE=true
- Eliminate Terraform Error Message about empty Authority Field
- Ensure instance id is set in logs on AWS when not using mesh.
- Execute Callback for empty HTTP request vector
- make num_chaff_requests not have a static lower bound
- Make terraform for aws delete cloud maps without error
- Redirect misleading log to /dev/null
- scorecard.yaml version updates
- shuffle request order (real and chaff) on SFE
- Update setup_2 and demo terraform configs to valid state
- Upgrades google terraform plugin to 5.31.0 to fix crash
- [IsolateBuyerAndSellerCodeExecution] Add wrapper for buyer's reportWin() udf
- [reporting] Enable reportResult execution for topLevelSeller
- Add chaffing feature flags in SFE/BFE.
- add EventMessage to log context
- Add request creation timestamp to ProtectedAuctionInput
- Add request/response to ExecuteInternal on async clients
- Add tee-container-log-redirect option in terraform
- Bash script for ASG and Cloud Map Custom HealthChecks
- Consented request replace enable_adtech_code_logging in Bidding Server
- Create inference model store for model management
- Create logging library for inference consented logs
- Debug Reporting for Bid Currency
- decrease aws/build_and_test duration by ~75%
- Feature Flag for TLS in Service Mesh
- Force the ML model reset with the probability of 0.1%
- Forward per request consented debugging config to roma callback
- Implement AWS Cloud Un-Map
- Implement new SFE <> BFE request format for chaffing
- Implement the probabilistic model reset for PyTorch
- Implement the probabilistic model reset for TensorFlow.
- Load Test Flag for AWS
- Log consented debugging information in inference sidecar
- log EventMessage in servers
- log udf log in EventMessage for non_prod debug_info
- Remove enableAdtechCodeLogging flag value
- Remove Envoy Access Logging
- Route consented inference requests to a consented model store
- send chaff requests from SFE
- Service Mesh in AWS
- Support CPU isolation in the inference sidecar
- update code/cloud build to use tags
- Upgrade AWS Provider for Terraform from v3.xx to v4.xx
- Use gRPC for AWS Service Mesh Envoy HCs
- [IsolateBuyerAndSellerCodeExecution] Add a new class for buyer's reporting code fetch and load
- [IsolateBuyerAndSellerCodeExecution] Add a new code wrapper with only scoreAd and reportResult
- [IsolateBuyerAndSellerCodeExecution] Add config flag to enable seller and buyer code isolation
- [IsolateBuyerAndSellerCodeExecution] Modify seller_udf_manager to fetch and load buyer udfs
- [IsolateBuyerAndSellerCodeExecution] Refactor the code fetch files.
- add check back in default grpc client
- Add DebugInfo pointer for debugging log into RomaRequestContext
- Add generation_id to chaff requests
- clean up log verbosity 3
- do not try to impersonate service accounts if TEST_MODE=true
- Eliminate Terraform Error Message about empty Authority Field
- Ensure instance id is set in logs on AWS when not using mesh.
- Execute Callback for empty HTTP request vector
- make num_chaff_requests not have a static lower bound
- scorecard.yaml version updates
- Upgrades google terraform plugin to 5.31.0 to fix crash
- Add tee-container-log-redirect option in terraform
- Create logging library for inference consented logs
- Forward per request consented debugging config to roma callback
- Support CPU isolation in the inference sidecar
- [IsolateBuyerAndSellerCodeExecution] Add a new code wrapper with only scoreAd and reportResult
- clean up log verbosity 3
- Add the CPU affinity lib for inference
- Cap debug reporting URL
- Fix the recency value passed to generateBid and scoreAds
- make aws-image-tag optional
- remove unitOverride in gcp seller dashboard terraform
- replace ABSL_LOG with PS_LOG in reactors
- update empty device signals to be an object, not string
- [Debug Reporting] Enable debug reporting by default
- [DebugReporting] Add a log to get the libcurl queue time
- add ability to tag AMIs
- Add kv e2e tests environment
- Add markdown-link-check to pre-commit
- add support for AWS CodeBuild
- add support for GCP Cloud Build
- API Updates for Bid Currency Support for Top-Level Seller in Server-Orchestrated Multi-Seller Auctions
- Enable TCMalloc for all B&A servers
- support sending Nitro Enclave logs to AWS CloudWatch for debugging
- [Temp] Adds buyer bid to top level auctions for Chrome < v125
- Adjust ps_vlog pre-commit hook to avoid recursing symlinks
- clean up log verbosity 2
- deps: Upgrade clang-format to v18
- deps: Upgrade pre-commit hooks
- Ensure that the debug url length is capped
- Add runtime check for correct inference backend version
- Adds debug info to secure invoke response
- Adds implementation of GetComponentAuctionCiphertexts API
- Adds optional flag to configure cloud platforms for component sellers
- allow AWS Session Manager instance connection
- enable bucket fetching for score ad
- Enables and adds tests for GetComponentAuctionsCiphertexts API
- Enables Top Level Auctions feature
- Implement Support for Bid Currency in Reporting
- Set Seller Rejection Reason on AdWithBid for Mismatch with Expected Buyer Currency
- Set Seller Rejection Reasons on AdScore for currency-related rejectons
- Support inference runtime configurations
- Support Interest Group Origin in InterestGroup and AuctionResponse
- Top Level auction does not accept multiple results from same seller
- upload aws/gcp hashes as artifacts on release
- Correct check condition in inference main
- Increase the gRPC recv message limit for the inference sidecar
- Pass complete ad object to scoreAd
- Remove redundant clang-tidy repo, move config to .clang-tidy
- RPC is finished properly in top level auction with reporting
- Temporarily add bazel_clang_tidy repo
- Add Bid Currency to PAS bidMetadata for scoreAd()
- Add blob fetch library
- Adds API for generating ciphertexts for server component auctions
- Adds proto mapping functions for top level auction reactor
- Adds reactor for handling top level auction in SFE
- Adds utility for re-encrypting ciphertexts from device
- Adds validator functions for top level auction reactor
- Build version dynamically for telemetry
- Check Bid Currency on Protected App Signals AdWithBids
- create SellerCodeFetchManager for auction service UDF fetching logic
- Enable cpp_no_warn mode by default
- Enable ML inference in B&A
- Implements server component auctions feature in SFE
- Implements support for server component auctions in secure invoke
- Implements top level auction feature in auction service
- Implements top level auctions support in secure invoke
- Add get_workspace_status script to github repo
- Create extra function to only add zone attribute to gcp deployment
- Delete Right AdWithBids for Currency Mismatch in SFE
- Do not check trailing space for patch files
- improve graceful shutdown of bidding server
- Simplify V8Dispatcher::LoadSync, switch to absl::Notification
- Update data plane library dependency
- update generate bid to receive empty json object when no device signals are present
- deps: Upgrade build-system to 0.57.0
- create BuyerCodeFetchManager to handle buyer udf fetching logic
- enable consented debugging in AWS deployment
- integrate buyer code fetch management
- Output raw metric for consented request
- Add a metric attribute for instance region
- Add bid rejection reasons for bid currency
- Add bid_currency to CBOR Auction Response
- Add binary HTTP Utils
- add component_reporting_urls to SelectAdResponse
- Add KV client for TEE retrieval
- Add pylint to pre-commit
- Add TF flags for PAS KV/retrieval service
- Adds API support for server component auctions
- Adds HPKE encryption utility for server component auction
- Adds OHTTP Encryption utils for sharing with top level auction reactor
- Adds proto field to support top level auction
- Adds util for resolving AuctionScope with SelectAdRequest
- Check bid currency on AdWithBids against buyer currency in SFE
- Disable/enable core dumps based on build flavour
- Enable prettier pre-commit hook for JavaScript
- Enable prettier pre-commit hook for JSON
- Instantiate KV GRPC Client in bidding
- Integrate bidding with KV service
- Oblivious HTTP Utils
- PAS contextual ads API changes
- Pass Contextual PAS Ads from SFE -> BFE
- Relay contextual PAS Ads from BFE -> Bidding
- Support Bid Currency in the ScoreAdsReactor
- update gcp collector to use internal proxy network load
- Use GRPC client for ad retrieval
- Validate currencies in the auction config
- Avoid race conditions when turning PyTorch models into eval mode
- correct branching condition in register model request
- Correct git submodule links
- Fix debug statements to include result status
- Fix flaky reporting_helper_test
- resolve inference sidecar path for unit test
- update kv service hash after repo name change
- Fix minor typos in load testing doc
- deps: Downgrade build-system to v0.52.0
- deps: Upgrade build-system to 0.53.0
- deps: Upgrade build-system to 0.55.1
- deps: Upgrade build-system to 0.56.0
- deps: Upgrade functionaltest-system to v0.12.0
- [reporting] Add set modifiedBid value in sellerReportingSignals for multi seller auctions
- Add metric for size of protected_ciphertext and auction_config
- add additional metrics to the AWS seller dashboard
- Add AWS perfgate benchmark
- Add aws s3 support to perf-test-helpers
- Add metric attribute for instance zone
- Add perfgate analyzer support
- Add Support for Buyer Experiment Group ID
- Add tf vars for otel collector template file and image uri
- API Changes for Bid Currency
- Builds AWS AMI with debug otel collector
- create AWS seller and buyer metric dashboards for monitoring
- load and make inference request with a PyTorch model
- Seller KV Experiment Group ID
- Test Bid Currency on AdWithBids returned from generateBid()
- ad metadata string is escaped to be valid json in browser
- Add run_all_tests bazel config
- Bazel builds should ignore the cost tool
- Consider all 2XX HTTP codes as success
- deps: Upgrade data-plane-shared-libraries to 44d1d64 2024-01-08
- Ensure that non-200 status codes become errors
- Fix build flag and copybara rule
- Fix custom quickstore input file bug
- Fix debug statements to include result status
- Fix flaky sandbox_executor_test
- Prevent Config from being copied into RomaService
- Removes a test breaking the build
- secure_invoke client stub lifetime prolonged to wait for call to end
- Temporarily revert changes to the SUT so that it passes
- turn on metric noise (a minimum of 5 QPS is recommended for noised metric)
- require consented token min length 6
- Add metric attribute for B&A release version
- Add metric attribute for operator name
- Add perf-test-helpers library
- add sfe metric for request with win ad
- component auction: Adds test for component auction support in secure invoke
- dep: Update build-system to release-0.52.0
- Import perfgate exporter and uploader tars
- require consented token min length 6
- turn on metric noise (a minimum of 5 QPS is recommended for noised metric)
- Update perf environment to use custom Otel Collector
- Use hardcoded adtech code and kv mock from e2e-testing env in perf env
- Auction service parses adMetadata object
- Consider non-positive desirability ads as rejected
- Fixes broken test
- Log JS errors conditionally
- Make the error message compatible with deterministic CBOR
- minimize secure_invoke reliance on default arguments
- Remove rejected ads from consideration in scoring
- Update B&A to integrate RomaService Changes
- [reporting] Add noiser and bucketer for noising reporting inputs
- [reporting] Pass modeling signals through noiser before being input to reportWin
- Add ps verbosity tf var
- Add scoring support for PAS
- Add support for new OHTTP request format.
- build PyTorch from source for B&A inference server.
- Collector script for performance testing
- component auction: Auction server passes top-level seller to scoreAd
- component auction: Auction service parses output and skips allowComponentAuctions = false ads
- component auction: Bidding server passes top-level seller to generateBid
- component auction: Bidding service parses output and skips allowComponentAuctions = false bids
- component auction: BuyerFrontEnd Service accepts and forwards top level seller
- component auction: Return error for Android device orchestrated component auctions
- component auction: SellerFrontEnd service accepts and forwards top level seller
- component auction: SFE service parses output cbor encodes for chrome
- includes git info tag for gcp docker images
- monitor key fetch metrics for B&A servers
- Partition request fail metric by the status message
- Upgrade functionaltest-system to v0.11.0
- Add check for correct key ID in select_ad_reactor tests.
- consented logger memory leak
- Fixes test failures in select ad app reactor
- log missing key id in grpc status
- match copybara strip style for PyTorch build from source
- refactor gcp artifact build scripts to be more modular
- Remove local platform for common repo and update CloudPlatform
- Remove unnecessary reporting flags in gcp config
- update GCP build to write image_digest, not image id
- deps: Upgrade data-plane-shared to commit 115edb3 2023-11-16
- add debug info in sfe response
- Add reporting support for PAS
- Instrument KV request and response sizes
- Parameterize the collector startup script
- write docker sha256 image digest to file for GCP releases
- Send chaff when no desirable ad is found
- Add common error count by error code as a sever monitoring metric
- deps: Upgrade build-system to 0.49.1
- Add support for interestGroupNames
- Temporarily Hardcode Maximum Buyers Called to Two
- Use set instead of vector for Key-Value Server input key fields
- Change Set type to bTree to ensure deterministic key order
- read PS_VERBOSITY from cloud parameter
- Update visibility target to public target in common repo
- api: Add fields for component auction support
- Add load testing guide
- support passing in a custom public key/id to secure_invoke
- [reporting] Fix the description for component_seller_reporting_urls and top_level_seller_reporting_urls in the API
- [reporting] Remove unnecessary flags in aws config
- increases limit for default gRPC message size between servers
- deps: Upgrade build-system to 0.48.0
- Add GLOG env vars to local functional testing auction service container
- Add PAS support in bidding service
- add ps_verbosity flag to bidding, auction
- pass a client_type url param to KV services only when
- run otel collector inside a docker image on Container Optimized OS
- Skip call to Bidding Server when no Trusted Bidding Signals present
- update the service local start scripts to use docker
- Fixes crash in SFE/BFE due to curl handle lookup failure
- update functional test with new client type enum
- Add more buyer flags
- add support for a dedicated non-tls healthcheck port
- add utilization in cpu metric
- allow granular GCP per-region machine config
- change GCP sfe and bfe health checks to use grpc
- simplify GCP SFE envoy config
- Add Ad Retrieval flags for AWS
- Disable use of secure_invoke for functional testing
- Ensure the easy handle is copied
- Fixes the metadata keys sent to bidding
- Use the verbose level 1 for consented logs
- deps: Upgrade build-system to 0.45.0
- add boundary for unsafe histogram metric
- Add docker-based sut using data from fledge sandbox
- Add HTTP async client for ads retrieval
- Add MultiLogger as an unified logging interface
- add noise attribute to metric
- Add secure_invoke rpc invoker for rpc diff testing
- Allow clients to pass in a PublicKeyFetcher into KeyFetcherFactory.
- Allow services to override the code blob versions
- Deserialize ads retrieval response
- Log plaintext responses in B&A servers
- Log plaintext ScoreAdsRawRequest in Auction
- Remove require-ascii pre-commit check on changelog
- Skip call to Auction Service and return chaff when no scoring signals present
- Take auction type in account in GenerateBid
- Upgrade to functionaltest-system 0.9.0
- Add a missing header file
- Fix a broken test in MultiLogger
- Revert the change in builders and testing/functionaltest-system
- Add C++ code coverage support
- Add PUT support in HTTP client
- Add SetContext() in ConsentedDebuggingLogger
- Change secure_invoke output to be easily parsable by automated scripts
- Log plaintext GenerateBidsRawRequest in Bidding
- override metric's default public partition
- PAS support for BFE
- turn on external load balancer logging by default
- Change generate_bid_code_version from int to string in API
- [logging] Check for log fields in the response json from reporting scripts
- [logging] VLOG the logs, errors and warnings exported from Roma for reporting.
- [reporting] Add ad_cost and interest_group_name inputs to the reportWin function input
- Add copyright line in license block
- Add missing adCost to buyer's browser signals in the code wrapper
- Allow ScoreAd to return a number
- Check rapidjson property presence before access
- Fix invalidCode error when buyer's script is not a correct expression
- Fixes a comment in proto
- revert documentation update until the build_and_test_all_in_docker script is updated
- safe metric should not LogUnsafe
- Set the log context in BFE
- Updates the client types comments
- Change generate_bid_code_version from int to string in API
- Add code blob flags for bidding/auction
- Log plaintext GetBidsRawRequest in BFE
- PAS support in SFE
- Propagate consented debug config from BFE to Bidding
- Propagate consented debug config from SFE to BFE & Auction
- [reporting] Change the order of cbor encoding for win reporting urls in the response.
- [reporting] Populate win reporting urls for seller in top_level_seller_reporting_urls instead of component_seller_reporting_urls
- [reporting] Remove unnecessary signalsForWinner in the response from reporting wrapper function
- Disable enableSellerDebugUrlGeneration flag by default
- Add B&A App Install API
- Add ConsentedDebuggingLogger to write logs via OTel
- Add feature flag enabling/disabling PAS
- Add IG origin to AuctionResult
- add max surge to instance groups
- Add owner field for PAS Ad with bid
- Adds docker build option for local testing
- Adds PAS buyer input to GetBids
- Check debug token in ConsentedDebuggingLogger
- Include IG owner in buyer input
- Log decoded buyer inputs for consented debugging in SFE
- make export interval configurable
- make metric list configurable
- OpenTelemetry logging can be disabled via TELEMETRY_CONFIG flag
- Propagate Ad type from Auction => SFE
- update GCP terraforms to apply updates to instances without
- update LB policy to default to RATE instead of UTILIZATION
- Upgrade to functionaltest-system 0.8.0
- use private meter provider not shared with Otel api
- Bump required tf to 1.2.3 to be compatible with replace_triggered_by
- Change componentAds -> components
- Correct GCP dashboards
- Do not set runtime flags with empty strings
- Fixes race condition in SFE reactor due to mutex lock
- Log a message in case of server flag lookup failures
- Removes callback execution from default_async_grpc_client to prevent double invocation
- specify default cpu util before sending requests to other regions
- update OTel demo documentation
- Use bazel config to set instance and platform
- deps: Upgrade build-system to 0.43.0
- if your generateBid() returns ad component render urls in a field named "adComponentRender". You will have to update this to return these in a field named "adComponents". Changed name in the API. Updated in reactors. Added a unit test.
Testing: Unit test Bug: b/294917906 Change-Id: Ie4344f55b18ef10f7a81b197ec997be393fa7368
- Adds Readme for running B&A servers locally
- Enable CBOR encode/decode for ConsentedDebugConfig
- implement dp aggregated histogram metric
- include local server start scripts
- periodic bucket fetching using BlobStorageClient
- Correct paths to terraform modules in AWS demo.tf files
- Improve clarity of aws vs gcp cases
- Improve flag error handling
- include required dep for bucket code fetcher
- Remove enable_buyer_code_wrapper flag.
- Remove enable_seller_code_wrapper flag.
- Remove enableSellerCodeWrapper flag from aws and demo configs
- remove unnecessary flags in terraform configs
- Set cbuild flag --seccomp-unconfined
- Change AdWithBid.ad_component_render to .ad_components to align with OnDevice generateBid() spec
- Update terraform comments to communicate requirement for env name <= 3 characters
- [reporting] Add helper function for cbor encoding and decoding of Win Reporting Urls
- [reporting] Execute reportWin script and return the urls in the response.
- [reporting] Fix reportWin function execution.
- add AWS enclave PCR0 in release notes
- Add bazel configs for roma legacy vs sandboxed
- Add OSSF Scorecard badge to top-level README
- Add OSSF Scorecard GitHub workflow
- Add the server flags to support user consented debugging
- change autoscale cpu utilization measure to GCP suggested default (60%)
- clarify default lb policy within an instance group is ROUND_ROBIN
- Enable logging in the OTel collector
- Encrypt GetBidsRequest for benchmarking
- Flag protect the opentelemetry based logging
- move observable system metric to server definition
- Upgrade data-plane-shared-libraries for Opentelemetry logs API
- Encode Ad Component Render URLs when sending request to Seller Key-Value Server
- Fixes few ASAN issues
- Minimal CBOR encoding for uint/float
- Order the keys in the response
- Remove unwanted seller_origin flag from start_auction script
- Rename the opentelemetry endpoint flag
- secure_invoke respects --insecure flag for BFE
- api: Add the fields to support adtech-consented debugging
- deps: Upgrade build-system to v0.39.0
- deps: Upgrade build-system to v0.41.0
- deps: upgrade opentelemetry-cpp to 1.9.1 for B&A servers
- update release README instructions
- Update secure_invoke README with new instructions for running.
- [reporting] Add helper function to build the dispatch request for
- add buyerReportWinJsUrls to terraform files and enable_report_win_url_generation to BuyerCodeFetchConfig
- add cpu memory metric for debugging
- add GCP metric dashboard
- add method to accumulate metric values before logging
- changing PeriodicCodeFetcher to add wasm support and runtime config support
- load and append wasm string to the JS code wrapper
- support different instance types per service in GCP
- Upgrade build-system to release-0.31.0
- Add seller_code_fetch_config and buyer_code_fetch_config to server start scripts
- CPU monitoring is not limited to a specific environment
- Define container image log variables once
- Don't end select ad request prematurely
- patch google dp library to replace logging with absl
- Update and read the buyer_bids_ with lock
- deps: Upgrade build-system to v0.33.0
- add FetchUrls utility to MultiCurlHttpFetcher
- enable counter metrics with dp aggregation
- update multi-region.tf to use prod js urls and test mode true
- use //:non_prod_build to configure build
- Adjust all/small test configs
- Adjust sanitizer configs
- Correct example BFE hostname
- Correct license block
- Ensure --gcp-image flags are specified for gcp
- Ensure --instance flag is specified
- fix missing telemetry flag and readme
- Improve build_and_test_all_in_docker usage text
- [AWS] add example terraform directory with README
- [GCP] add example terraform directory with README
- Add bazel build flag --announce_rc
- add build_flavor for AWS packaging
- add build_flavor for packaging
- include coordinator and attestation support for GCP
- Upgrade build-system to release-0.30.1
- Adjust SFE DCHECKs
- bidding_service_test
- Change PeriodicCodeFetcher to use std::string instead of absl::string_view in the parameters
- refactor the test to share initialization
- remove unnecessary flags
- TEE<>TEE fix
- temporarily eliminate requirement to have device signals to generate bids
- add VLOG(2) for code loaded into Roma
- Correct xtrace prompt
- Add --instance to build_and_test_all_in_docker
- Add bazel configurations for platform and instance flags
- Add flag to config telemetry
- Add smalltests kokoro config
- changing MultiCurlHttpFetcherAsync to take a raw pointer executor.
- create a header file for PeriodicCodeFetcher object wtih constructor, necessary functions, and necessary variables
- create a source, BUILD, and test files for PeriodicCodeFetcher.
- create CodeFetcherInterface as an interface for different CodeFetcher classes.
- enforce list order in metric definition
- implement dp to aggregte (partitioned)counter metric with noise
- Implement GCP packaging scripts (including SFE envoy
- Implement Metric API used by B&A server
- Implement Metric context map
- Implement Metric router to pass safe metric to OTel
- Limit build_and_test_all_in_docker to run small tests only
- modify auction_main.cc and bidding_main.cc to integrate PeriodicCodeFetcher for code blob fetching.
- move GCP instances behind a NAT
- reactor own metric context
- remove hardcoded scoreAd.js and generateBid.js
- update TrustedServerConfigClient to work with GCP
- use telemetry flag to configure metric and trace
- Use terraform flag to specify if debug or prod confidential compute
- [GCP] add docker redirect flag for prod images
- [GCP] specify port instead of port_name for lb healthchecks
- Add bazel coverage support
- add GCP as a bazel config
- add missing gcp flags
- auction service returns if no dispatch requests found
- BFE Segfault after grpc reactor Finish
- complete removal of sideload IG data
- do not reference ScoreAdsReactor private members after grpc::Finish
- flaky auction and bidding tests
- GCP SFE dependencies were outdated
- gcp terraform should use env variable for buyer address
- mark docker and async grpc client stub tests large
- potential SFE segfault due to logging after Finish
- Remove --without-shared-cache flag
- rename service to operator, component to service
- Replace glog with absl_log in status macro
- skip scoring signal fetch and Auction if no bids are returned
- Specify test size
- test: Add size to cc_test targets
- update gcp packaging script to support all repos globally
- update GCP SFE runtime flags with new values
- update init_server_basic_script to use operator
- update managed instance group healthcheck
- update pending bids so SFE does not hang
- Use --config=value in .bazelrc
- Use buf --config flag
- validate that the buyer list is not empty
- build: Emit test timeout warnings from bazel
- build: Add scope-based sections in release notes
- build: Adjust small-tests configuration
- build: Create all-tests configuration
- Initial release