Highlights
- Tool result now adds external link
- Support runbooks by alerts & symptoms
- Improvements for llm context usage
- Improved Coralogix & Datadog toolsets
What's Changed
- Fix formatting of task status check example by @zyxue in #1174
- [ROB-2722] - Coralogix improvments by @Avi-Robusta in #1180
- ROB-2761 Avoid using model list file in specific tests by @RoiGlinik in #1183
- Document custom_args for Bedrock 1M context window by @aantn in #1182
- fix: correct venv installation path in dockerfile by @mainred in #1181
- Toolsets: Improved datadog/traces toolset and fixed datadog/rds prerequiste check by @Sheeproid in #1178
- [ROB-0000] Coralogix docs update by @Avi-Robusta in #1192
- remove the datadog/rds plugin as datadog/metrics plugin covers it by @Sheeproid in #1194
- feat: add healthz and readyz endpoint for pod probe by @mainred in #1187
- fixes env vars to make eval work with openai models by @zyxue in #1175
- [ROB-2769] Links toolsets by @Avi-Robusta in #1190
- enables datadog logs and traces to use pagination cursor correctly by @Sheeproid in #1199
- Reduced output tokens for multiple tool calls by using minified json by @Sheeproid in #1197
- Always handle grafana dashboards urls using the grafana dashboards to… by @arikalon1 in #1198
- Fixed broken link and added page to nav by @pavangudiwada in #1201
- Add timeout to kafka prerequisites check by @aantn in #713
- ROB-2713 improve healthcheck for tempo and loki by @RoiGlinik in #1189
- Add runbook alerts to prompt by @moshemorad in #1205
- Add image link comments to docker workflows by @aantn in #1210
- verify setup success in eval 64_keda_vs_hpa_confusio and change to hard by @Sheeproid in #1223
- 57_wrong_namespace fix judge to allow get logs from the right namespace by @Sheeproid in #1222
- 82_pod_anti_affinity_conflict: ensure anti-affinity rule exist in setup by @Sheeproid in #1221
- improving several tests setup and instructions by @Sheeproid in #1227
- stabilizing more tests by @Sheeproid in #1233
- Add retention warning to dev image PR comments by @aantn in #1224
- fixing failures due to non standard pod name in eval 101. test 100b is for testing nonstandard labels. by @Sheeproid in #1237
- ad 176_network_policy_blocking_traffic_no_runbooks based on test 84 by @Sheeproid in #1236
- faster reliable setup for some counting and Kubernetes tests by @Sheeproid in #1238
- faster setup for 162_get_runbooks by @Sheeproid in #1239
- ROB-2714 more default crd permissions by @RoiGlinik in #1204
- helm: add oci release for charts by @DrFaust92 in #1188
- docs: use concrete Anthropic model example instead of placeholder by @aantn in #1235
- add a regression test tag that is to be run in PRs instead of easy tests by @Sheeproid in #1241
Full Changelog: 0.17.0...0.18.0-alpha