8000 Release v1.5.0 · openziti/ziti · GitHub
[go: up one dir, main page]
More Web Proxy on the site http://driver.im/
Skip to content

v1.5.0

Pre-release
Pre-release
Compare
Choose a tag to compare
@github-actions github-actions released this 25 Mar 14:05
· 167 commits to main since this release
v1.5.0
This tag was signed with the committer’s verified signature.
plorenz Paul Lorenz
GPG key ID: 1A4AC953FAF6AD08
Verified
Learn about vigilant mode.
695b077
This commit was created on GitHub.com and signed with GitHub’s verified signature.
GPG key ID: B5690EEEBB952194
Verified
Learn about vigilant mode.

Release 1.5.0

What's New

  • Bug fixes and features
  • Change to router endpoints file default name
  • Updated Cluster Defaults
  • Updates to terminator costing
  • Router metrics changes

Router Endpoints File

The router endpoints file used to have a default name of endpoints.
As it is a YAML file, the file now has a default name of endpoints.yml.
This may affect existing setups running the beta HA code. If this is
a concern, the endpoints file path can be configured:

ctrl:
  endpointsFile: /path/to/endpoints.file

Updated Cluster Defaults

The following defaults have changed.

cluster:
  # How many updates before creating a new snapshot. 
  #
  # New default: 500
  # Old default: 8192
  snapshotThreshold: 500 

  # How many old entries to keep around, so that a stream of
  # entries can be sent to sync peers, instead of sending an
  # entire snapshot
  #
  # New default: 500
  # Old default: 10240
  trailingLogs: 500

Terminator Costing Changes

When a terminator is selected for a service dial, and that dial results in a failure, a failure cost
is added to that terminator. This will bias future dials towards other terminators, if they are available.

The failure cost can be reduced by successful dials. Failure cost is also reduced over time. In previous
releases this was a fixed credit of 5, every minute. This is now changing to an exponential
amount, based on time since the last failure.

If X is minutes since last failure, the credit is now: min(10, 2 ^ (X/5)).

Router Metrics Changes

There are four new router metrics, focused on visibility into flow control.

  • xgress.blocked_by_local_window_rate - meter which ticks whenever an xgress becomes blocked by the local window being full
  • xgress.blocked_by_remote_window_rate - meter which ticks whenever an xgress becomes blocked by the remote receive buffer being full
  • xgress.blocked_time - timer which tracks how long xgresses are in a blocked state.
  • xgress_edge.long_data_queue_time - timer that tracks times to process incoming data payloads to xgress_edge.

The xgress_edge.long_data_queue_time will be controller by a router config file setting. It will default to disabled. The other metrics will always be enabled.

Router metrics has two new configuration setting:

metrics:
  # Number of usage events to be able to queue. Defaults to 256. If this queue backs up, it can
  # slow down dispatch of data from an SDK onto the fabric.
  eventQueueSize: 256

  # If set to true, enables the xgress_edge.long_data_queue_time metric. Defaults to false.
  enableDataDelayMetric: false

Component Updates and Bug Fixes

Assets 11
Loading
0