Highlights from NetApp Insight 2018’s: 1265-2 Troubleshooting Your Infrastructure with OnCommand Insight

NetApp Insight 2018 - Recorded Insight Sessions

OCI & Extensible Schema

- OCI 7.3.1 and higher allow for the OCI “acquisition” team to extend existing schema objects with new counters or attributes:

--- Counter = time series data, i.e you can plot it on a chart.

--- Attributes = a point in time fact. If the most recent data source poll collects a new value, it replaces the old.

- OCI acquisition team can do this via Data Source Service Packs or even data source patches.

- This allows us to address some feature requests without requiring an actual OCI full or maintenance release.

- In OCI 7.3.1+ all .patch files contain a metadata.json which defines the extensions.

The Benefit of the Extensible Schema

- OCI 7.3.1+, users can obtain these new to OCI metrics by installing a Service Pack, which is much faster to install than performing an actual OCI upgrade.

- Expect more of this and please stay on top of Service Packs.

- You have-the-ability to write data into OCI via some REST API calls under the term “Integrations”.

Image: OCI REST API integrations

Integrations Rules of Engagement

- If you want visibility into facts or time series data, and you are willing to collect and parse the data, writing that into OCI may allow you to eliminate needing to run other products for visualization.

- If you want alarming - not yet. OCI 7.3.x does not support configuring performance policies on integration data.

- If you want reporting in the DWH on this data, you might be over architecting a solution. The OCI DWH does have a dwh_custom database you can write to. It may make sense to just write the data into that directly.

But I Lost My “Dummie’s Guide to Python”

- Perhaps you don’t have the time or desire to invest in learning Python, or an equivalent, to facilitate feeding data into OCI.

- Or, perhaps you just want to see what you can do with the data in OCI, before investing in scripting…

What is collectd?

- A pluggable, lightweight open source infrastructure for collecting and routing performance metrics.

- Plugins are both available for bringing data in and sending data out.

See:

collectd – The system statistics collection daemon

https://collectd.org/

- But collectd does NOT natively let you visualize data - you need to send it into another stack.

- Enter OCI’s collectd write plugin:

--- https://github.com/NetApp/OCI_collectd

--- Use the collectd ecosystem to collect data, visualize it in OCI.

--- The collectd plugin needs a token from OCI that you can generate by making an authenticated REST API query to the OCI …/rest/ integrationAgents call.

--- Collectd data will flow in as integration data, available for dashboard usage.

Working with Integration Data

- Collectd data will appear as integration data available for selection in custom dashboards beneath all of the standard OCI inventory objects.

- The names of the counters will reflect what plugins you are using in collectd AND have chosen to route to OCI with your collectd.conf.