0.0.12

Quick bugfix for ui accessing routes with access token middleware

0.0.11: Better logging, simplified configs, better state

• Moved towards better, structured logging
• It is no longer necessary to specify critical-alerter, warning-alerter or warning-threshold, critical-threshold. If you do not want to alert on some condition - you don't have to (ie. if you would prefer to read the state from the API instead). Or in other words, if you do not want to have any alert conditions on warning - you can avoid adding any warning threshold or alerter; same for critical.
• State entries are now set to auto-expire within 2*check-interval length; this was an oversight from the getgo - state entries should disappear (especially if the check is no longer running). If the interval cannot be determined (bug?), state entries will expire in 24 hours

Changelog: 0.0.10...0.0.11

0.0.10: Overwatch related bugfixes

This release contains a number of fairly critical bugfixes that are related to the overwatch component release.

- Fixed #80
- Fixed case where the runMemberMonitor did not properly shutdown when Stop() was called in cluster
- Fixed case where overwatch would stop accepting messages (and thus blocking calling messagers)
- Fixed various bits in cluster.go that prematurely exited their loopers

0.0.9: This is a fairly beefy update. Thanks to all the contributors!

* Check configuration can be fully managed via the API now (#139)
    * [Refer to the API readme docs](https://github.com/9corp/9volt/tree/master/docs/api)
* `9volt` API now supports token auth (if enabled) (#141)
* `9volt` is now much more resilient when `etcd` failures occur (#140)
    * Introduced the `overwatch` component that is responsible for managing internal components such as `alerter`, `cluster`, `director` etc during etcd failure(s)
    * Read more about it [Here](https://github.com/9corp/9volt/blob/master/docs/OVERWATCH.md)

0.0.8: Minor alerter message change

- Updated alerter to include check description in messages

0.0.7: Not too shabby

This release contains a bunch of fixes and improvements from various folks:

• Improvements for etcd failures
• An mkdir -p-like behavior for dal.Set()
• A sane replacement for the spaghetti incident that is known as base-monitor.go (ie. use fsm VS a bunch of conditionals)
• Dedupe resolve messages
• Authenticated etcd support

AND it looks like docker image building + push is now working via Travis!

docker pull 9corp/9volt:0.0.7 or docker pull 9corp/9volt.

Fixes and minor improvements

Various fixes:

• Introduce 'semver' along with 'version'; available in member status, /version endpoint and --version via cli.
• critical state bugfix #126 ; thanks @jessedearing
• more logging #124
• adding ssl certs into the docker image (apk add ca-certificates), otherwise https:// checks will fail
• Vendored handlers #125

Updated env vars

This release includes a very minor env var related change: #118

Docker image updated and pushed as 0.0.5 (docker pull 9corp/9volt).

Tags, DNS check, bugfixes

Another alpha release, this one contains a number of fixes and new things:

• Check-pinning via node and check tags is now fully working ( #103 )
• DNS check introduced by @relistan ( #109 )
• Internal check-statistic bug fix ( #103 )
• Multiple node tag fix ( #115 )
• Tests + build now ran on Go 1.8 ( #113 )
• Official 9volt docker images are now in DockerHub: docker pull 9corp/9volt
• docker-compose updates
  • docker-compose up will now spin up a 3 node cluster (1 with no tags, 1 with foo tag, 1 with foo and bar tags)
  • This is helpful for getting a feel for what 9volt is all about

Have fun!

Check distribution and leader election bugfixes

It appears that leader election was semi-busted; fixed in #104.
Check (re)distribution was semi-busted as well; fixed in #102.

Note: Picking members for new checks logic still needs to be made 'tag' aware. Should get updated/fixed soon.

Woo!