About Metraylze

For advanced digital teams, Metralyze is the time-series analytics platform that can spot cause-and-effect relationships across multiple data sources. Unlike traditional monitoring and analytics tools, Metralyze isn't confined to data managed inside a single tool, but can tap multiple data sources across IT, CX and finance.

Metralyze is a cloud-based service.

What is time-series data?

Any data with a timestamp, basically. Time-series data usually tracks how one or more metrics develop over time, with additional tags or dimensions giving additional context to the metrics. In a fast-changing world, nearly all data is time-series data.

What's so great about time-series data?

In years of working with IT Operations teams that manage customer-facing systems (such as E-Commerce shops, booking portals, SaaS products), we have seen the same set of problems over and over again:

1. People don't have time to proactively look at monitoring data, except when customers report issues. Only a fraction of the data is incorporated into dashboards, and only a fraction of the dashboards are maintained and get attention.
2. Threshold-based alerts don't work: there's always too many or too few alerts.
3. Monitoring data is spread over various systems, and hardly anyone can traverse several layers for a root-cause analysis. But root causes are always connected: problems in one network partition caused a service to become partially unavailable, which caused a different service to run out of threads, which caused errors to be shown to customers.
4. When there's a root cause analysis done, it consists of screenshots from monitoring systems, and it's hard to follow the same path of investigation again if the problem reoccurs.

At the same time, there's huge potential in time-series monitoring data - to generate early warnings of customer impact, to find possible scalability bottlenecks, to make capacity/cost calculations, and many more use cases.

How does Metralyze solve those problems?

Metralyze connects to various sources of time-series information, from monitoring tools and APMs up to web analytics systems. By continuously requesting time series data from those systems at different levels of aggregation, Metralyze can find patterns in data from different sources. From simple threshold-based alerts to detecting change points and specific patterns in time-series, the product architecture allows using the same analysis algorithms on multiple data sources.