v1.2.1

This is a security only release.

Upgrade third-party modules in response to security issues:

• urllib3 1.24.2 (CVE-2019-11324)
• Jinja2 2.10.1 (CVE-2019-10906)

Install and upgrade guides at https://knightjoel.github.io/zpark/

v1.2.0

• Upgrade third-party modules in response to security issues
  • Requests 2.2.20 (CVE-2018-18074)
  • urllib 1.23 (CVE-2018-20060)
  • Flask 0.12.4 (CVE-2018-1000656)
• Improved the output of "show status" when Zpark is talking to Zabbix 3.4+.
  Stats better reflect what's shown in the Zabbix dashboard now. Note that
  the count of disabled triggers doesn't match due to some behavioral
  differences between the data returned by the Zabbix API vs the Zabbix
  client/server protocol. There's not much that Zpark can do about that.
• New setting ZABBIX_TLS_CERT_VERIFY. Set this to False if your Zabbix
  installation uses a self-signed certificate. Defaults to True.
• In order to aid troubleshooting, Zpark will emit a log entry when it
  receives a command from a Spark user that isn't on the trusted users list.
  This message is emitted at log level DEBUG which can be configured by
  modifying the API_LOG_HANDLER setting.

API_LOG_HANDLER = {
        'class': 'logging.handlers.SysLogHandler',
        'address': '/dev/log',
        'facility': 'local6',
        'level': 'DEBUG'
}

You shouldn't have logging set to DEBUG unless you're troubleshooting. It
creates an opportunity for someone to fill your log file with their arbitrary
commands.

Install and upgrade guides at https://knightjoel.github.io/zpark/

v1.1.0

• The SPARK_TRUSTED_USERS config parameter now supports specifying
  @domain.com to make it easy to whitelist users from an entire organization.
  Eg: SPARK_TRUSTED_USERS = ['@example.com', '[email protected]']
• Allow [;:,] as a delimeter between the bot's name and the command in
  @-mention messages. Eg: @bot, show issues
• New command: hello. The bot will respond with an introduction about itself
  including a list of supported commands. Note the bot will only respond to
  hello if the user is on the list of trusted users.
• A much better job is done to determine the bot's own name when it receives
  a message via @-mention. This allows the bot to have a more complex name
  and to respond to @-mention commands when there is another user in the room
  with a similar name to the bot.

Install and upgrade guides at https://knightjoel.github.io/zpark/

Initial release!

Features:

• Relays new Zabbix alerts (which are generated from Zabbix events) to Spark
  user(s) and/or room(s).
• Relays Zabbix "all clear" alerts (which are generated when a recovery
  condition is met) to Spark user(s) and/or room(s).
• Configure Spark message recipients based on a combination of alert
  severity, time of day, host group, or other arbitrary criteria. eg:
  • Send all alerts to the Spark room "Network Engineers" between 0800-2000 Mon-Fri
  • Send alerts with severity “critical” or higher to engineers Joe, Jill, and Mark 24x7, using their Spark accounts "[email protected]", "[email protected]" and "[email protected]".
• Immediate notification of alerts. As soon as the bot receives an alert
  from Zabbix, it is immediately sent to Spark; there is no queueing or
  batching of messages.
• Interrogate information from Zabbix by issuing commands to the bot
  on Spark. Eg, "show status".
• Decentralized: run the bot on the Zabbix server or on a separate
  server, your choice.

Install and configuration guide at https://knightjoel.github.io/zpark/