Skip to content

Commit

Permalink
prepare v2.1.0 release
Browse files Browse the repository at this point in the history
  • Loading branch information
FZambia committed Dec 30, 2018
1 parent 34c8b82 commit 5388aad
Show file tree
Hide file tree
Showing 4 changed files with 47 additions and 6 deletions.
22 changes: 22 additions & 0 deletions CHANGELOG.md
Original file line number Diff line number Diff line change
@@ -1,3 +1,25 @@
v2.1.0
======

This release contains changes in metric paths exported to Graphite, you may need to fix your dashboard when upgrading. Otherwise everything in backwards compatible.

Improvements:

* Refactored export to Graphite, you can now control aggregation interval using `graphite_interval` option (in seconds), during refactoring some magical path transformation was removed so now we have more predictable path generation. Though Graphite paths changed with this refactoring
* Web interface rewritten using modern Javascript stack - latest React, Webpack instead of Gulp, ES6 syntax
* Aggregated metrics also added to `info` command reply. This makes it possible to look at metrics in admin panel too when calling `info` command
* More options can be set over environment variables – see [#254](https://github.com/centrifugal/centrifugo/issues/254)
* Healthcheck endpoint – see [#252](https://github.com/centrifugal/centrifugo/issues/252)
* New important chapter in docs – [integration guide](https://centrifugal.github.io/centrifugo/guide/)
* Support setting `api_key` when using `Deploy on Heroku` button
* Better timeout handling in Redis engine – client timeout is now bigger than default `redis_read_timeout` so application can more reliably handle errors

Fixes:

* Dockerfile had no correct `WORKDIR` set so it was only possible to use absolute config file path, now this is fixed in [this commit](https://github.com/centrifugal/centrifugo/commit/08be85223aa849d9996c16971f9d049125ade50c)
* Show node version in admin web panel
* Fix possible goroutine leak on client connection close, [commit](https://github.com/centrifugal/centrifuge/commit/a70909c2a2677932fcef0910525ea9497ff9acf2)

v2.0.2
======

Expand Down
13 changes: 12 additions & 1 deletion docs/content/server/configuration.md
Original file line number Diff line number Diff line change
Expand Up @@ -358,14 +358,25 @@ http://localhost:8000/debug/pprof/

– will show you useful info about internal state of Centrifugo instance. This info is especially helpful when troubleshooting. See [wiki page](https://github.com/centrifugal/centrifugo/wiki/Investigating-performance-issues) for more info.

#### Healthcheck endpoint

New in v2.1.0

Use `health` boolean option (by default `false`) to enable healthcheck endpoint which will be available on path `/health`. Also available over command-line flag:

```bash
./centrifugo -c config.json --health
```

#### Custom internal ports

We strongly recommend to not expose API, admin, debug and prometheus endpoints to Internet. The following Centrifugo endpoints are considered internal:

* API endpoint (`/api`) - for HTTP API requests
* Admin web interface endpoints (`/`, `/admin/auth`, `/admin/api`) - used by web interface
* Prometheus endpoint (`/metrics`) - used for exposing server metrics in Prometheus format
* debug endpoints (`/debug/pprof`) - used to inspect internal server state
* Healthcheck endpoint (`/health`) - used to do healthchecks
* Debug endpoints (`/debug/pprof`) - used to inspect internal server state

It's a good practice to protect those endpoints with firewall. For example you can do this in `location` section of Nginx configuration.

Expand Down
2 changes: 1 addition & 1 deletion docs/content/server/monitoring.md
Original file line number Diff line number Diff line change
Expand Up @@ -33,4 +33,4 @@ To enable automatic export to Graphite (via TCP):

By default stats will be aggregated over 10 seconds interval inside Centrifugo and then pushed to Graphite over TCP connection.

If you need to change this aggregation interval use `graphite_interval` option (in seconds, default `10`).
If you need to change this aggregation interval use `graphite_interval` option (in seconds, default `10`). This option available since v2.1.0
16 changes: 12 additions & 4 deletions misc/release/notes.md
Original file line number Diff line number Diff line change
@@ -1,10 +1,18 @@
**Important** If you are using `rpm` or `deb` packages from packagecloud.io then you have to re-run the [installation method of your choice](https://packagecloud.io/FZambia/centrifugo/install) for Centrifugo repository. This is required to update GPG key used. This is a standard process that all packages hosted on packagecloud should do.
This release contains changes in metric paths exported to Graphite, you may need to fix your dashboard when upgrading. Otherwise everything in backwards compatible.

Improvements:

* Redis TLS connection support - see [issue in Centrifuge lib](https://github.com/centrifugal/centrifuge/issues/23) and [updated docs](https://centrifugal.github.io/centrifugo/server/engines/#redis-engine)
* Do not send Authorization header in admin web interface when insecure admin mode enabled - helps to protect admin interface with basic authorization (see [#240](https://github.com/centrifugal/centrifugo/issues/240))
* Refactored export to Graphite, you can now control aggregation interval using `graphite_interval` option (in seconds), during refactoring some magical path transformation was removed so now we have more predictable path generation. Though Graphite paths changed with this refactoring
* Web interface rewritten using modern Javascript stack - latest React, Webpack instead of Gulp, ES6 syntax
* Aggregated metrics also added to `info` command reply. This makes it possible to look at metrics in admin panel too when calling `info` command
* More options can be set over environment variables – see [#254](https://github.com/centrifugal/centrifugo/issues/254)
* Healthcheck endpoint – see [#252](https://github.com/centrifugal/centrifugo/issues/252)
* New important chapter in docs – [integration guide](https://centrifugal.github.io/centrifugo/guide/)
* Support setting `api_key` when using `Deploy on Heroku` button
* Better timeout handling in Redis engine – client timeout is now bigger than default `redis_read_timeout` so application can more reliably handle errors

Fixes:

* Resubscribe only to shard subset of channels after reconnect to Redis ([issue](https://github.com/centrifugal/centrifuge/issues/25))
* Dockerfile had no correct `WORKDIR` set so it was only possible to use absolute config file path, now this is fixed in [this commit](https://github.com/centrifugal/centrifugo/commit/08be85223aa849d9996c16971f9d049125ade50c)
* Show node version in admin web panel
* Fix possible goroutine leak on client connection close, [commit](https://github.com/centrifugal/centrifuge/commit/a70909c2a2677932fcef0910525ea9497ff9acf2)

0 comments on commit 5388aad

Please sign in to comment.