v0.3.0-beta0

Caveats

This is a pre-production beta release. We will not introduce breaking changes to the API between minor versions, however, compatibility between beta versions is not guaranteed (e.g. v0.2.0-beta1 and v0.2.0-beta2).

Changes

• feat: Iris Ho: adding cronjob option for collector by @loranger2k in #286
• feat: implement atomic operations by @yuryu in #300
• feat: Checksums - proto by @yuryu in #307
• feat: Checksums - add Checksums and Digest types by @yuryu in #306
• feat: support CRC32C and MD5 checksums by @yuryu in #309
• feat: Optimistic locking for records by @yuryu in #320
• feat: support redis pool config from file by @lorangf in #325
• feat!: migrate to go-redis as redis client by @hongalex in #334
• feat!: add default expiration (5m) for Redis by @yuryu in #335
• feat: add support for inequality filters by @hongalex in #338
• feat: add support for ordering in QueryRecords by @hongalex in #340
• feat: query limit support by @hongalex in #343
• feat: chunk count checking by @yuryu in #361
• feat: make max inline blob size configurable by @ano12ak in #375
• feat: support keys_only in QueryRecords by @yuryu in #369
• feat: add grpc health service by @hongalex in #382
• feat: bundle grpc_health_probe in image by @hongalex in #387
• feat: support graceful shutdown on sigint/sigterm by @hongalex in #385
• feat: enable opencensus tracing with cloud trace for datastore by @hongalex in #400
• feat: add new commit with upload method by @hongalex in #402
• feat: support query offset by @loranger2k in #409
• feat: get multi records by keys by @loranger2k in #410
• feat: use google.rpc.Status by @yuryu in #411
• feat: add configurations for keep alive on grpc server by @vasconcelosvcd in #416
• feat: add SignedUrls on every chunk in gcs by @vasconcelosvcd in #417
• fix: revert deletion of generated proto files by @loranger2k in #418
• feat: replace OpenCensus telemetry with OpenTelemetry by @Piotr-MPG in #423
• feat: tracing propagation and service name prioritize envvar by @vasconcelosvcd in #427
• feat: expose new redis config by @vasconcelosvcd in #429

Bug fixes

• fix: return chunk size and count with GetRecord by @yuryu in #312
• docs: add key terms and table of contents by @karenarialin in #308
• fix: wrong CRC32C checksums error message by @vasconcelosvcd in #317
• fix: remove defer statement closing listeners by @hongalex in #388
• fix: Server shutdown grace period by @loranger2k in #390
• fix: clean go proto in Makefile properly by @hongalex in #392
• fix: backward compatibility with numberofchunks by @zaratsian in #393
• fix!: remove indexing for timestamps by @hongalex in #399
• fix: remove extra writes when uploading chunked data by @loranger2k in #405

Full Changelog: v0.2.0-beta2...v0.3.0-beta0

v0.2.0-beta2

Caveats

This is a pre-production beta release. We will not introduce breaking changes to the API between minor versions, however, compatibility between beta versions is not guaranteed (e.g. v0.2.0-beta1 and v0.2.0-beta2).

Changes

This release is primarily a hot fix to address a crashing bug.

• feat: remove opaque string limit check
• refactor(pubsub): replace gob with msgpack
• fix: crash in CreateChunkedBlob with invalid param
• docs: Tweaking steps for garbage collection

v0.2.0-beta1

Caveats

This is a pre-production beta release. We will not introduce breaking changes to the API between minor versions, however, compatibility between beta versions is not guaranteed (e.g. v0.2.0-beta1 and v0.2.0-beta2).

New features

• Chunk blob support
• Opaque string support
• Lots of docs improvements
• Updated dependencies

For all changes, please check the v0.2.0-beta1 milestone on GitHub.

Docker images are available as

• gcr.io/triton-for-games-dev/open-saves-server:v0.2.0-beta1
• gcr.io/triton-for-games-dev/open-saves-collector:v0.2.0-beta1

v0.1.0-beta0

Caveats

This is a pre-production beta release. We will not introduce breaking changes to the API, however,

Breaking changes

• The Blob field in the Blob message was removed and replaced with the new streaming blob apis
• The default command line options for project and bucket names were removed
• Some fields were removed from the Datastore entities and the new version cannot read the old entities
• The primary git branch has been renamed from master to main
• The project has been renamed from Triton to Open Saves

New features

• Streaming blob support
• Log level is now configurable
• New deployment guide
• Garbage collector for blob entries
• Cloud Run support
• New sample client

For all changes, please check the beta0 milestone on GitHub.

v0.1.0-alpha2

Features

• feat(cache)!: add support for cache hints (#123
• feat(cache)!: use cachestore to cache request in server (#111)
• feat: Add streaming methods to BlobStore (#120)
• feat: Add a state machine to Blob (#101)
• feat(config): Env var config (#88) (#122)

Fixes

• fix: make create/update methods return new resource (#112)
• fix: keep old CreatedAt value in UpdateRecord

Known Issues

• There is no large blob or cache support yet.
• The HTTP REST endpoint is not implemented.
• GKE support is not available yet; deploying to Cloud Run is the only option at the time.

v0.0.0-alpha1

Caveats

This is a pre-production alpha release and you should expect breaking API changes before we release v1.0.
Cloud Datastore entities created in v0.0.0-alpha0 is not compatible with alpha1. No migration path will be provided.

What's New

• Cloud Run deployment scripts
• Timestamps (created_at and updated_at) for Store and Record.
• Ping method

Known issues

There is no large blob or cache support yet
The HTTP REST endpoint is not tested and the current implementation will be removed from future releases. We are discussing better options to support REST APIs in the future.

v0.0.0-alpha0

Caveats

This is a pre-production alpha release and you should expect breaking API changes before we release v1.0.

What's new

• Basic metadata operations: CreateStore, CreateRecord, DeleteRecord, GetStore, ListStores, DeleteStore, GetRecord, UpdateRecord
• GCP Datastore support for the backend

Known issues

• There is no large blob or cache support yet
• The HTTP REST endpoint is not tested and the current implementation will be removed from future releases. We are discussing better options to support REST APIs in the future.