Originally published at: https://www.influxdata.com/blog/release-announcement-telegraf-1-7-3-influxdb-1-6-1-and-kapacitor-1-5-1/
A new maintenance release for Telegraf is available now.
- Telegraf 1.7.3 — release notes
- The output buffer is now locked when adding new metrics to avoid concurrency issues.
- The agent will preserve metric type when using filters in output plugins.
- The required docker API version has been reduced to 1.21.
- Updated the error message to be more descriptive when the url cannot be parsed.
- We now skip bad entries for this input.
- Preserve leading whitespace commonly used for indentation when logging stack traces.
- Fixed a parse error that happened on FreeBSD 11.2.
- InfluxDB 1.6.1 — release notes
- Performance Improvements
- Allows for faster merging of large series sets. This improves query planning time by approximately 20%; however, a large amount of time is still spent building the bitsets.
- Significantly improve the performance of
influx_inspect buildtsi. This makes building a TSI index around 4-10 times faster than previous releases.
- Improve the performance of the storage read service when querying a single measurement by providing a more efficient means to get the field keys for a measurement and using an iterator that only iterates over that single measurement, rather than iterating over all measurements.
- Configuration Options
- Security — existing behavior is maintained by default; however, if you wish to restrict the security options by TLS version and a specific list of supported ciphers, the ability to do so is now made possible through configuration settings.
- Compactions — by default the maximum number of processors used by the compaction service represents 50% of the total with no upper limit. This restores the behavior prior to the 1.5 release. You may set a fixed limit to a number smaller than 50% of the available processors.
For our InfluxEnterprise customers, log in to the InfluxEnterprise portal and download the latest binaries from there.