Welcome to Trickster 2.0.5! 🎉

In this release, we:

• added support for more aggregation operators in Time Series Merge: limitk, limit_ratio, stddev, stdvar and quantile and sort
• support OTLP tracing protocol selection between http (default) and gRPC
• added max_query_range config to time series backends
• support TSM Replica Groups for indicating HA replicas vs disjoint merge members
• added a new listeners section to the config to define all frontend listeners in one place. You can now configure multiple frontend listener ports and map each backend to a listener. Existing listener configs in mgmt and frontend will work with a deprecation warning.
• incorporated various other performance improvements and bug fixes
• improved handling of merge errors

Thanks to:

• @houyuwushang
• @thinker0
• @LimitlessEarth

# via ghcr.io
docker run --name trickster -d -v /path/to/trickster.yaml:/etc/trickster/trickster.yaml -p 0.0.0.0:8480:8480 ghcr.io/trickstercache/trickster:2.0.5

# via docker.io
docker run --name trickster -d -v /path/to/trickster.yaml:/etc/trickster/trickster.yaml -p 0.0.0.0:8480:8480 docker.io/trickstercache/trickster:2.0.5

helm install trickster oci://ghcr.io/trickstercache/charts/trickster --version '^2' --set image.tag="2.0.5"

For more information, see:

• Trying Out Trickster
• trickster's helm chart.

What's Changed

• docs: update OTLP tracing examples by @houyuwushang in #1049
• feat: support OTLP tracing protocol selection by @houyuwushang in #1050
• refactor: remove unused tsm rr fallback by @houyuwushang in #1054
• docs: document X-Trickster-Result header by @houyuwushang in #1055
• fix: close demand probe response bodies by @houyuwushang in #1057
• fix: return 500 on internal failures by @jranson in #1053
• fix: ensure host header promotion works consistently by @jranson in #1058
• fix: batch high-cardinality tsm rank merges by @houyuwushang in #1060
• test: modernize benchmarks with b.Loop by @houyuwushang in #1065
• fix: merge aggregations through prometheus sort wrappers by @houyuwushang in #1064
• feat: configure backend CORS policies by @houyuwushang in #1066
• fix: enable TSM context cancellation by @jranson in #1062
• test: remove healthcheck wall-clock waits by @houyuwushang in #1074
• refactor: introduce TSM merge plans by @houyuwushang in #1075
• cleanup string literals + reorganize by @jranson in #1076
• test: update Prometheus SDK integration calls by @houyuwushang in #1085
• feat: support limit_ratio in ALB TSM by @houyuwushang in #1077
• sec+feat: harden config endpoint and add listeners map by @jranson in #1086
• support prom scalars in ALB Time Series Merge by @jranson in #1082
• fix(routing): swap middleware wrapping order of WithResourcesContext and LimitQueryRange by @thinker0 in #1078
• feat(alb,proxy): enforce max_query_range limit on ALB TSM scatter/gather entry point by @thinker0 in #1079
• chore: standardize on timeconv.Duration across options (issue #1089) by @thinker0 in #1090
• fix: preserve duration parsing compatibility by @houyuwushang in #1092
• feat: support TSM HA, Federation and Hybrid via Replica Groups by @jranson in #1088
• chore: integration cleanup for linter/coverage by @jranson in #1093
• tests: increase test coverage in least-covered packages by @jranson in #1094
• tests: add more unit tests by @jranson in #1095
• feat: support pooled stddev and stdvar in ALB TSM by @houyuwushang in #1080
• bugfix: retain errorType after replica group merge by @jranson in #1096
• feat: support exact quantile in ALB TSM by @houyuwushang in #1083
• tests: increase daemon test coverage by @jranson in #1097
• feat: support global PromQL limitk in ALB TSM by @houyuwushang in #1084

Full Changelog: v2.0.4...v2.0.5