v1.1.53

What's Changed

• chore: update production workflow to publish upgrade artifacts V2 by @abc3 in #336
• chore: enable nodelay in db_handler by @abc3 in #340

Full Changelog: v1.1.51...v1.1.53

v1.1.51

What's Changed

• chore: redefine funs on upgrade by @abc3 in #323
• chore: filter poolboy_sup during the upgrade by @abc3 in #325
• docs: update development docs by @J0 in #329
• feat: add communication via partisan for client and db handlers interaction by @abc3 in #317
• feat: auth error caused by invalid characters in user or db_name by @abc3 in #334
• fix: handle invalid secret formats in parse_secret by @J0 in #326

Full Changelog: v1.1.45...v1.1.51

v1.1.45

What's Changed

• chore: update runtime.exs path in hot_upgrade by @abc3 in #322

Full Changelog: v1.1.44...v1.1.45

v1.1.44

What's Changed

• fix: set Logger level otherwise Phoenix plug exceptions are not thrown by @chasers in #312
• fix: async calls for all ready_for_query messages by @abc3 in #318
• fix: stop pool connection if credentials fail by @abc3 in #321

Full Changelog: v1.1.40...v1.1.44

v1.1.40

What's Changed

• fix: sync sending after reaching the threshold by @abc3 in #313

Full Changelog: v1.1.39...v1.1.40

v1.1.39

What's Changed

• fix: enable connection to custom database name other than 'postgres' by @w3b6x9 in #254
• fix: tenant id handling in manager by @abc3 in #271
• feat: add new metrics by @abc3 in #270
• feat: pick node to spin up a pool by tenant id by @abc3 in #269
• fix: metrics tenant by @abc3 in #272
• fix: start_dist in client_handler by @abc3 in #275
• fix: kill the linked handler in session mode by @abc3 in #277
• feat: add more logs by @abc3 in #278
• chore: set max heap size for handlers by @abc3 in #280
• fix: setting a mode on transaction port by @abc3 in #283
• feat: setting log level per tenant via options by @abc3 in #286
• feat: update hostname by @abc3 in #288
• chore(deps): bump shlex from 1.2.0 to 1.3.0 in /native/pgparser by @dependabot in #281
• chore: improve logging by @abc3 in #292
• chore: set number of ets tables and processes by @abc3 in #293
• fix: add [active: once] on receiving a response by @abc3 in #295
• feat: reset the check timer after the client's disconnection by @abc3 in #291
• fix: handle client termination in db handler by @abc3 in #296
• fix: add an error case for get_payload() by @abc3 in #297
• fix: stop db handler before client by @abc3 in #298
• feat: local or remote pool checkout duration metric by @abc3 in #299
• feat: recheck auth secrets on connection error by @abc3 in #304
• feat: idle workers in pool by @abc3 in #285
• chore: ecto pool_size 25 by @abc3 in #305
• chore: update options for the get secrets by @abc3 in #306
• feat: clear tenant's cache on update or delete by @abc3 in #309

New Contributors

• @w3b6x9 made their first contribution in #254

Full Changelog: v1.1.13...v1.1.39

v1.1.13

Notable

• Fixes "expected SASL response, got message type" which can appear in Postgres logs
• Don't attempt to check database availability during tenant CRUD operations
• Fixes metrics per tenant endpoint
• Fixes edge cases during scram auth

What's Changed

• feat: track duration from tcp connection to greeting clients by @abc3 in #258
• fix: don't block the connection during the scram dance by @abc3 in #256
• fix: add a new line in tenant metrics by @abc3 in #261
• fix: new line in tenant metrics by @abc3 in #262
• feat: bump Rustler and pg_query to resolve security warning by @J0 in #255
• fix: add proto guards by @abc3 in #267
• feat: don't check db creds on update by @abc3 in #265

Full Changelog: v1.1.6...v1.1.13

v1.1.6

Notable

• Fixes a bug which caused prepared statements to fail in session mode

What's Changed

• fix: linking db pid in session mode by @abc3 in #249

Full Changelog: v1.1.5...v1.1.6

v1.1.5

Notable

• Pools now start with only 10 connections and create new ones up to tenant default_pool_size or user pool_size
• Logs correct tenant id
• API endpoints accept PATCH requests

What's Changed

• fix: DbHandler Logger metadata by @abc3 in #245
• fix: update api by @abc3 in #247
• feat: set initial pool size to 10, enable other via max_overflow by @abc3 in #225

Full Changelog: v1.1.2...v1.1.5

v1.1.2

Notable

• allow_list field on the tenant to support network restrictions
• More docs
• client_heartbeat_interval on the tenant to detect zombie client connections
• Bug fixes
• Observability improvements

What's Changed

• feat: reduce logging by @J0 in #216
• Update pgbouncer.md (fix typo) by @zorbash in #224
• fix: return proper error on undefined tenant by @J0 in #220
• chore(deps): bump rustix from 0.38.10 to 0.38.28 in /native/pgparser by @dependabot in #217
• chore: upgrade gh runner by @abc3 in #226
• feat: add heartbeat interval to client handler by @abc3 in #223
• fix: remove stub in heartbeat_interval by @abc3 in #229
• fix: catch errors from decode_payload by @abc3 in #231
• feat: update the Dockerfile and install Rust so build goes through by @J0 in #237
• fix: higher res buckets for pool checkout duration metric by @chasers in #241
• fix: config compile_time_purge_matching for prod by @chasers in #242
• feat: check ip against allow list by @chasers in #240
• fix: adds a FAQ by @chasers in #235
• chore(deps): bump tj-actions/changed-files from 35 to 41 in /.github/workflows by @dependabot in #243
• feat: update dockerfile and add docker-compose file by @J0 in #215
• feat: add additional clause for Telem.client_join by @J0 in #219

New Contributors

• @dependabot made their first contribution in #217

Full Changelog: v1.0.0...v1.1.2