You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
* WIP: Helm support for v1/v2
* Make helm deployable again
* Milestone: v2 micros-services works
* Fix folders and search addressed to fix single-binary v2
* Update helm-docs to support ignored
Also install it using go get instead docker
* Build helm always without CGO
* Introduce cluster label change when switching from single binary to
micro-services
Without that there is a chance the single binary instances are kept in
the ring without being cleaned up.
* Provide info about the migration in the NOTES
* Add some WIP v2 migration docs for helm
* Update migration guide
* Upgrade helm-ci
* Update image for super linter
* Fix linting issues
* Call helm-docs correctly
* Fix lint check
* Make superlinter work
* Not too sure super linter brings any value
* Apply suggestions from code review
Co-authored-by: Alberto <[email protected]>
* Apply suggestions from review everywhere
---------
Co-authored-by: Alberto <[email protected]>
Once data before Step 2 is no longer relevant, we can get rid of the v1 components. This will loose all data before Step 2.
36
+
37
+
```
38
+
helm upgrade \
39
+
pyroscope \
40
+
./operations/pyroscope/helm/pyroscope \
41
+
--set architecture.storage.v1=false \
42
+
--set architecture.storage.v2=true \
43
+
--set pyroscope.persistence.enabled=true
44
+
```
45
+
46
+
47
+
## Micro-Services
48
+
49
+
### First deploy v1
50
+
51
+
- Note: Needs persistence enabled otherwise data will be lost after restart
52
+
- Note: `--set architecture.overwriteResources.requests.cpu=10m` allow this to be tested without allocation many resources, this should not be used in production
| architecture.deployUnifiedServices | bool |`false`| Deploy unified write/read services. These endpoints will can be used no matter if the helm chart is configured as single-binary or microservices |
29
+
| architecture.microservices.clusterLabelSuffix | string |`"-micro-services"`| Memberlist cluster label that will be used for all members of this cluster |
30
+
| architecture.microservices.enabled | bool |`false`| Enable micro-services deployment mode. This is recommend for larger scale deployment and allow right size each aspect of Pyroscope. |
31
+
| architecture.overwriteResources | object |`{}`| This flag is useful for testing, it will overwrite all pods resource statements with its contents |
32
+
| architecture.storage.migration.ingesterWeight | float |`1`| Specifies the fraction [0:1] that should be send to the v1 write path / ingester in combined mode. 0 means no traffics is sent to ingester. 1 means 100% of requests are sent to ingester. |
33
+
| architecture.storage.migration.queryBackend | bool |`true`| Specify a time stamp from when the v2 read path should serve traffic. |
34
+
| architecture.storage.migration.queryBackendFrom | string |`"auto"`| Specify a time stamp from when the v2 read path should serve traffic. |
35
+
| architecture.storage.migration.segmentWriterWeight | float |`1`| Specifies the fraction [0:1] that should be send to the v2 write path / segment-writer in combined mode. 0 means no traffics is sent to segment-writer. 1 means 100% of requests are sent to segment-writer. |
| serviceMonitor.targetLabels | list |`[]`| ServiceMonitor will add labels from the service to the Prometheus metric https://github.com/prometheus-operator/prometheus-operator/blob/main/Documentation/api.md#servicemonitorspec|
98
112
| serviceMonitor.tlsConfig | string |`nil`| ServiceMonitor will use these tlsConfig settings to make the health check requests |
99
113
100
-
----------------------------------------------
101
-
Autogenerated from chart metadata using [helm-docs v1.8.1](https://github.com/norwoodj/helm-docs/releases/v1.8.1)
0 commit comments