Monitoring & Observability 2019 Predictions

Observability’s buzzwordiness will remain, but the solutions are going to get more interesting and widespread. A lot of existing monitoring vendors are already working on implementing tracing features, and I expect we’ll start to see more from them in the area of high cardinality data support. By the end of 2019, I think there will be much less of a feature gap between Honeycomb/LightStep (who lead the way in this area) and the more traditional vendors. The differentiating factors are likely to be around usability rather than raw capability.

Serverless monitoring tools are going to disappear. Well, sort of. They’re going to get consumed by the larger, all-in-one vendors, which makes sense, as monitoring a serverless infrastructure isn’t substantially different from monitoring, say, containers, and the container monitoring market was also consumed by the larger vendors once the approach was settled and the market penetration/hype of containers was sufficient to warrant the development time. Serverless is going to be there in 2019.

Incident management and response will be the next frontier. With VictorOps and OpsGenie acquisitions in 2018, and many well-known and respected people talking much more about human factors (eg, Allspaw, J. Paul Reed, Jason Hand, Jessica DeVita, and many more I’m sure I’ve missed), it seems inevitable that we’ll see more vendors start working on more robust incident management features. I think we’ll also start to see some brand new competitors in that market. Speaking of which, the major player in the incident management space, PagerDuty, is probably headed for an IPO in 2019. TechCrunch agrees, too.

Scalable, on-prem logging architectural patterns will become a thing. Grafana Labs is leading the way with the release of Loki, attempting to solve this problem: high-volume, easy-to-maintain, scalable, on-prem logging. Because “log less” is painful and not-actually-useful advice, and the existing patterns to solve this require a fair amount of glue (eg, Kafka + Elasticsearch) and a non-trivial amount of work, I think we’ll start to see more progress made in 2019 towards codifying patterns and turning it into something easier to deploy.

Disclosure: No vendor mentioned in this article has a financial relationship with Monitoring Weekly at the time of publication.