InfluxData's Q&A Session for Oct 13: Intro to InfluxDB Telegraf and the Plugin Ecosystem

By Support InfluxDB / Use Cases, Developer
Oct 14, 2016

We have posted the video archive for the Oct 13: Intro to InfluxDB Telegraf and the Plugin Ecosystem and the following is a recap of the Q&A session during the training.

Q: Is Raspberry Pi supported? I had a problem with memory issues using InfluxDB version 0.8.

A: We have completely changed the product and build pipeline since v0.8. We now have ARM support. In addition, if you’re looking to upgrade to v1.0 from v0.8 I recommend checking out one of our technical papers which goes into detail about the differences between the versions and the migration steps.

Q: In the beginning, you mentioned that timestamps can automatically be assigned.  Does Telegraf assign a timestamp when it receives each event? Or does this only happen when InfluxDB receives it?

A: InfluxDB Telegraf will assign timestamps as well.

Q: How would you recommend ensuring High Availability of each Telegraf agent? Specifically, regarding the host.

A: Load balancing an array of Telegraf instances works well for this. On a host level the systemd unit file that comes with the process should be sufficient. Please note that currently,  Telegraf does not support High Availability by itself. The best way to keep the agent up and running is through an external system. For example, you can run Telegraf inside a docker container and expose system resources/volumes to it so it can monitor host-level metrics.

Q: How does the exec process management work? I am familiar with collectd.

A: Exec will trigger your program and wait for it to produce output metrics until the configured timeout.

Q: So the exec plugin will reap children which don’t respond and restart them?

A: The plugin does not currently keep track of children processes. It is a good suggestion for future development.

Q: Are there any examples of running Telegraf in Docker - exposing the bits of the host OS to enable system monitoring?

A: We recommend that you look at published Docker guides that will help expose the required endpoints and volumes for system-level monitoring with Telegraf from inside the container.

Please make sure you sign up for our next free training session on “Intro to Kapacitor for Alerting and Anomaly Detection” webinar on 10/20