Last 7 days
0
Features: 0
Changes: 0
Fixes: 0
Deprecations: 0
Temporal
Durable execution platform and managed cloud for reliable distributed applications.
Latest Temporal changelog updates, official release notes, breaking changes, security patches, pricing changes, and developer reactions in one product feed.
Follow this Temporal release-notes page to spot useful features, risky migrations, noisy announcements, and source links before they hit your backlog.
Where Temporal updates are checked
Changes.Watch links back to official changelog and release-note sources so summaries stay easy to verify.
Related changelog channels
Use channels to follow groups of tools around a stack, workflow, or topic.
Temporal updates by type
Rolling windows show how many product updates landed in the last 7, 30, 90, and 365 days, grouped by existing changelog semantics.
0
0
5
25
Latest Temporal changelog updates
Upgrade base Alpine image to version 3.23.5.
Bumped default CLI version to 1.7.2 and upgraded Alpine base to 3.23.5.
Combined PostgreSQL visibility schema upgrades (v1.10‑v1.13) and added PGX SimpleProtocol tests with bigint duration encoding.
Updated core dependencies (x/crypto, x/net, Go, thrift) with a security patch for thrift
Updated default CLI to 1.7.0, upgraded Go toolchain to 1.26.4, thrift to 0.23.0, and applied security patches to CLI and crypto libraries.
Bumped dependencies in the v1.29.x branch
Bump project dependencies in the v1.30.x branch
Upgrade requires new core and visibility DB schemas (CHASM support, new tables and generated columns) – a breaking change for custom persistence implementations.
Potential breaking change: for replication setups using authorization, set system.disableStreamingAuthorizer=true to prevent connection errors.
Potential breaking change: replication setups using authorization must set system.disableStreamingAuthorizer=true to avoid connection errors.
Potential breaking change: replication setups using authorization must set system.disableStreamingAuthorizer=true to avoid connection errors.
Added a security check for namespaces in batch workflows (CVE‑2026‑5199, LOW severity).
Added security check for namespaces in batch workflow (CVE‑2026‑5199, low severity)
Bump dependencies to address security CVEs.
Bumped dependencies to address security CVEs.
Introduced Worker Versioning Upgrade on Continue as New (Public Preview) allowing pinned workflows to auto‑upgrade to the target worker deployment version
Updated Go dependencies (golang.org/x/crypto, golang.org/x/net) to address multiple CVEs
Updated dependencies to newer versions.
Docker images are now slimmed down for security, removing many tools and binaries, deprecating several images, and switching to embedded Sprig templating; Helm charts require a minimum version of 0.73.2, making this a breaking change.
Added backward‑compatible decoding of Worker Deployment memo by ignoring unknown fields during unmarshalling.
Patched critical vulnerabilities CVE‑2025‑14987 and CVE‑2025‑14986.
Patched security vulnerabilities CVE-2025-14987 and CVE-2025-14986.
Fixes for CVE-2025-14987 and CVE-2025-14986 vulnerabilities
Fixes an issue with Priority and Workflow Versioning
Breaking change: Docker images are slimmed down for security, removing binaries and packages such as the Temporal CLI, admin tools, curl, jq, and database clients.
Introduced a limit on the number of parts allowed in authentication tokens.
Limits the number of parts allowed for authentication tokens, preventing malformed or oversized tokens.
Fixes workflow update bugs
Required core schema upgrades (MySQL, PostgreSQL, Cassandra) and a breaking change for Worker Versioning deployments – you must delete and recreate deployments when moving from v1.27.x to v1.28.0.
Fix minor bugs in Worker Deployment functionality
Require core and visibility schema upgrades for MySQL/PostgreSQL and Elasticsearch before upgrading, with temporary performance impact.
Added TemporalPauseInfo column to visibility schemas (Elasticsearch, MySQL, PostgreSQL) – requires schema upgrade before cluster upgrade (breaking change).
Fixed minor bugs related to Nexus and Update with Start functionality.
Fixed Elasticsearch Scroll API bug where total hits tracking was disabled
Fix: disabled total hits tracking for Elasticsearch Scroll API
Introduce Nexus RPC framework and public preview of Workflow Update for durable cross‑namespace operations and mutable workflow updates
Added workflow lock acquisition during backfill history and support for NextRetryDelay option, preventing retry delays from being capped by MaximumInterval
Fixed the SQL visibility schema issue introduced in v1.24.0.
Standard visibility removed; upgrade to advanced visibility and update plugin names (mysql→mysql8, postgres→postgres12).
Updated server and dependencies to v1.23.1, addressing several security vulnerabilities.
Fixes a denial‑of‑service issue caused by invalid UTF‑8 data in the history service.
Fixes a denial-of-service vulnerability where malformed UTF-8 input to the history service could crash the server.
Mitigates a denial‑of‑service issue where malformed UTF‑8 data could crash the history service.
Replaced gogo/protobuf with google.protobuf, causing breaking changes to generated code and requiring core/visibility schema upgrades.
Fixed rollback issue introduced in v1.23.0 release candidates, with no impact on OSS users.
Ensures PollActivityTaskQueueResponse.ScheduleToCloseTimeout is always set, preventing nil dereference errors.
Updated server to version 1.22.4 and adjusted CI/CD setups to pull the version from go.mod
Updated golang.org/x/net to remediate three CVEs (security fix)
Added action metrics for schedule workflows, a force‑continue‑as‑new signal, and exposed buffer size/dropped‑action info in ScheduleInfo
Fix security issue in Temporal Docker image that could leak DB password via shell logs
Added experimental gRPC streaming API for cross‑cluster replication (disabled by default)
Fixed a security bug by removing debug logging of credentials from the Cassandra schema update tool
Reverts recent cache handling changes that caused excessive memory usage and blocked workflows under high load.
Fixed critical bugs causing workflow blocking under high load, BuildIDs attribute errors, and non‑deterministic errors in the Workflow Update feature.
Fixed visibility task generation for unchanged BuildIds, preventing errors in visibility processing.
Fix downgrade errors caused by new BuildIds search attribute incompatibility with 1.20 clusters
Fixed bugs: skip reading close event for retention tasks, use lower priority for standby tasks, added mutable state corruption protection, and corrected replication ack aggregation logic.
Introduces preview Worker Versioning with Build ID compatibility, requiring schema changes and new dynamic config flags.
Fix schedule and search attribute bugs, including cron handling and duplicate attribute types
Fixed schedule workflow bugs (timer reuse, initial memo timing, catch‑up window metric) and corrected search attribute handling
Fix several SQL visibility bugs and improve query converter handling, including pointer passing and deserialization errors
Removed SERVICE_FLAGS env var, matching.useOldRouting, and changed default authorizer behavior (system namespace and no‑namespace cases); OpenTelemetry counters now get a “total” suffix and Elasticsearch list support is deprecated, const...
Fixed multiple minor bugs in task processing and ES bulk processor (NPE, commit timeout, panic capture, serialization errors).
Upgrade storage schema versions (MySQL/Postgres to 1.9, Cassandra to 1.7) is required before moving to v1.19; server now runs on Go 1.19 and includes tctl v1.17.2.
Fixed history scavenger bug on delete mutable state
Fixed several minor bugs, including replication task nil handling and added retention validation in the history scavenger.
Fixed bug introduced in v1.18.2 that prevented server startup in a single-node configuration.
Fixed merge map payload handling and task reschedule policy
Removed Elasticsearch v6 support and added v8 support; upgrade ES first if on v6.
Fixed task‑queue processing failure when persistence is unavailable during load/reload
Fix bug in Retry Policy logic for Frontend Service handler
Fixes bugs in archival error handling, retry logic, and queue processor throttling
Fixed multiple metric and rate‑limiting bugs (tally timer, timer bucket, task/queue rate limiting, expired task dispatch) improving stability.
Fixed replication stack and shard lock scope bugs
Fixed replication stack issues and a potential deadlock scenario.
Fix workflow stuck issues during namespace failover and missing first task for abandoned child workflows
Fix duplicated callback listener ID in namespace registry and cluster metadata
Upgrade required schema versions for MySQL, PostgreSQL, and Cassandra with a new Schema Upgrade Tool and note that v1.16 is the last Alpine‑based Docker image (future releases need Docker >=20.10).
Fix potential shard deadlock issue (issue #2570)
Fixes potential shard deadlock issue (2570).
Added TLS support for multi‑cluster deployments, including separate server/client config and remote‑cluster certificate handling.
Add a new `tctl admin cluster list` subcommand and enable auto‑forwarding for all APIs; support local namespace promotion, global namespace updates, and removal of clusters from replication lists.
Fix for Golang security vulnerability CVE‑2021‑44716
Patched Golang CVE‑2021‑44716 vulnerability.
Fixed several minor bugs: version checker update, c‑symbol typo in caCert, invalid start‑time filter validation, and remote cluster addition handling.
Fixed history service bugs that could miss tasks during pod bounce and added trim logic for stale history events
Hotfix for Matching service to improve handling of many low‑traffic task queues.
Fix critical bug that could lose tasks when history pods bounce between nodes, urging upgrade to v1.14.3+.
Fix a critical bug that could lose tasks when history pods bounce between nodes, advising upgrade to v1.14.3+
Hotfix improves persistence stability under high load.
Critical bug fix preventing task loss; requires upgrading to patch 1.14.3 and running the schema upgrade (MySQL/PostgreSQL 1.7, Cassandra 1.6).
Fixes a hotfix issue with Elasticsearch document IDs exceeding 512 bytes, preventing errors for long Workflow IDs (>475 bytes).
Expose final statuses (Completed/Failed/TimedOut) for Cron and retried workflow executions, removing the previous “Continued As New” behavior (breaking).
Support for Elasticsearch v7 added; upgrade requires schema changes and a multi‑step migration, making it a breaking change.
Fixed Elasticsearch document ID length issue by capping IDs at 512 bytes
Hotfix applied to the matching engine for stability
Fix ignored InvalidArgument error for unknown custom search attributes when using custom mapper
Fix task queue rate calculation, TaskWriter panic on shutdown, and duplicate visibility task handling
Enhanced Visibility is now production‑ready with no schema changes.
Fixed cron schedule-related bugs
Fix blocked shard acquisition channel
Reverted PRs 1699 and 1697 to fix a deadlock issue in the matching service, restoring stability
Added reconnection policy configuration for gocql.ClusterObject
Fix workflow reset failure (issue 1673)
Fixed the issue preventing workflow resets (issue #1673).
Added configurable sync match wait time and dedicated metrics for context cancellation.
Upgrade dgrijalva/jwt-go to v4.0.0 preview1, fixing CVE‑2020‑26160 vulnerability.
Added namespace information to API error logs.
Introduce SearchAttributesManager and new tctl admin/cluster commands to add and list custom search attributes (add‑only, auto‑populate on server start).
Added upsert and delete visibility record support for the configured database only (issue #1532).
Added a startup check for TLS certificate expiration.
Breaking change: TLS hostname verification is now enabled by default for tctl, Cassandra and SQL tools; the old enable-flag is removed and a disable-flag is added to opt out.
Fixed a bug that could cause visibility data loss when using Cassandra persistence with infinite workflow timeouts
Fix potential data loss of visibility data when using Cassandra visibility persistence.
Fixed potential visibility data loss when using Cassandra visibility persistence with infinite workflow timeout (issue #1456).
Fixed bug where blob was mistakenly used as encoding type for DescribeMutableState API, correcting tctl admin workflow describe command and preventing error logs from Temporal worker
Fixed a bug where blob was incorrectly used as the encoding type for the DescribeMutableState API, preventing errors in tctl admin workflow describe and worker logs
Fixed bug where blob was incorrectly used as the encoding type for the DescribeMutableState API, eliminating errors in the tctl admin workflow describe command and worker error logs.
Fixed namespace tagging issue in matching and history service metrics.
Fixed a potential data race in the DescribeMutableState API (issue 1360)
• Breaking changes: Kafka dependency removed, dry‑run option in schema tools dropped, and metric naming changed (old "all" namespace will be removed in v1.10).
Fixed data race issues in the describe mutable state API and history service task processing.
Fixed data race in the DescribeMutableState API (issue 1360).
Fixed a Cassandra persistence bug that could corrupt workflow history.
Fixed a Cassandra persistence issue that could corrupt workflow history (Bugfix PR #1187).
Patch for Cassandra persistence issue that could corrupt workflow history
Fixes Cassandra persistence bug that could corrupt workflow history (PR #1187).
Fixes Cassandra persistence bug that could corrupt workflow history
Fixed a Cassandra persistence bug that could corrupt workflow history
Patched an issue in the non‑Kafka visibility queue implementation (see issue 1287).
Fixed issue with non‑Kafka visibility queue implementation.
Fixed server Docker image start-up failure when Elasticsearch is enabled but not configured
Fixed a bug in the Kafka deprecation logic introduced in the 1.6.x series.
Patched a bug in the Kafka deprecation logic introduced in previous versions
- Fixed a bug causing infinite timeout issues in the server runtime.
Fixed a bug causing infinite timeouts in the server.
- Fixed a bug in the infinite timeout feature
Deprecate Kafka for visibility and switch to internal visibility queues with dual‑processor mode for seamless migration.
Patched a bug in the new visibility queue processor affecting Elasticsearch/Kafka integrations.
Patched an issue with the infinite timeout feature.
Patch resolves infinite timeout issue in the feature
Deprecate Kafka visibility queue and replace it with an internal visibility queue, with detailed migration steps and config changes.
Fixed a bug in the infinite timeout feature, ensuring proper handling when timeouts are set to 0 (no timeout).
Updated persistence schema to version 1.2 – requires schema migration before server upgrade.
Patch released to resolve an issue present in v1.3.1.
Patched issue found in v1.2.1
Fixed issue present in 1.3.0
Moved metrics configuration to a global section and deprecated the per‑service metrics block
Added official support for PostgreSQL v9.6 and TLS configuration options for PostgreSQL and tctl.
Hotfix release v1.1.1 replaces v1.1.0
Introduce schema changes requiring migration using provided tools for MySQL and Cassandra
First production‑ready release of Temporal.
Fixed critical issue preventing the server from starting multiple times via Docker Compose (hotfix 740).
Non‑rollback release: data created in v0.30.0 is incompatible with older versions and will cause panics if downgraded.
Added a convenient docker‑compose file for MySQL and Elasticsearch deployment and new config options (public client, dynamic config editing, Cassandra timeout, configurable PublicClient in template.yaml).
Retry attempts now start at 1 and activities get a default retry policy, improving reliability out‑of‑the‑box.
Renamed import paths: go.temporal.io/temporal → go.temporal.io/api, go.temporal.io/sdk/temporal, go.temporal.io/server, and server/api packages.
Sets new default for ParentClosePolicy and removes executable permission from JSON files.
Introduce Failure/FailureInfo error types, track last RetryStatus, and enforce proper workflow timeout handling
Added TLS configuration support for Temporal clusters and updated payloads to include both headers and data fields.
Added TLS configuration support for Temporal (preparing tctl and web integration).
Last release that still includes Thrift endpoints for legacy client support.