0.2.0
NOTE: this release includes many breaking changes in the FlowCollector
CRD. We've tried to pack as many breaking changes as possible prior to releasing our GA (expected end of year). Some names / labels and the default namespace have changed, by renaming network-observability
to netobserv
. As a consequence, if you are upgrading from 0.1.4 or earlier, it is necessary to uninstall both the custom resource and the operator itself first.
Highlights
This release is placed under the sign of performance. There have been many improvements in that regard: in the eBPF agent, flowlogs-pipeline, Loki ingestion and querying, and with the Kafka-based processing pipeline. By the way, this Kafka pipeline is now a promoted feature (it was previously experimental).
Finally, we started an experimental feature in the console plugin: new overview dashboards. They don't have their definitive shape yet as we're still working on them, but you can have an early look if you wish: just add &overview=preview
in your browser address bar, when you're on a Console plugin page.
What's Changed
Breaking changes 🛠
- NETOBSERV-448: Moving agent type, ipfix and ebpf config into their own common subsection. by @mariomac in #162
- NETOBSERV-396: Flp prom tls by @OlivierCazade in #158
- NETOBSERV-448: reviewing API to follow conventions by @mariomac in #163
- NETOBSERV-555 NETOBSERV-552 CRD updates - many breaking changes by @jotak in #169
- NETOBSERV-660: Create metric section by @OlivierCazade in #181
Non-breaking changes
- updated labels used by prometheus for metrics produced by flp by @KalmanMeth in #150
- NETOBSERV-101 R&D: Kube enricher write path for downstream operator by @jpinsonneau in #135
- Add breaking-change triage in release notes by @jotak in #148
- Remove kafka experimental flag for next release by @jotak in #159
- NETOBSERV-101 - loki statusURL by @jpinsonneau in #154
- Improve release doc & script by @jotak in #151
- NETOBSERV-467 Updated defaults (eBPF / sampling), update doc by @jotak in #160
- Workflows update by @memodi in #157
- NETOBSERV-435 - Set default req/limits resources for eBPF agent by @jpinsonneau in #155
- Add default limits to CR by @jotak in #161
- NETOBSERV-497 Use direct-prom pipeline by @jotak in #153
- NETOBSERV-433 and NETOBSERV-373 - unify names by @jotak in #164
- Fix scc for sample-workload by @eranra in #166
- Allow enabling profiler port in FLP by @mariomac in #167
- Fix eBPF agent's sampling: 0 value to disabled by @mariomac in #168
- NETOBSERV-589: setup FLP transformer to receive protobuf for eBPF mode by @mariomac in #171
- NETOBSERV-579 Use netobserv prefix, add 4 new metrics, remove one by @jotak in #173
- Expose new Kafka parameters + tweak defaults by @jotak in #170
- Kafka mk targets by @memodi in #174
- Remove unused pipeline transfo and metrics by @jotak in #176
- NETOBSERV-489: Use authorize feature from console plugin API by @OlivierCazade in #175
- Fix auth token mode for FLP by @OlivierCazade in #178
- NETOBSERV-578: allow passing env vars to FLP by @mariomac in #180
- NETOBSERV-466: Add metric to count packet number in FLP by @OlivierCazade in #179
- NETOBSERV-447: Changed default FLP retry value to prevent FLP beeing oom-killed by @OlivierCazade in #182
New Contributors
Full Changelog: 0.1.4...0.2.0