You rarely know ahead of time what combination of attributes will be correlated with some degenerate behavior. It could be a bad build ID or a particular combination of customer + replica set.
Store all the context without the cost of slow queries or impossible aggregations. Don’t get stuck having to predict the future.
Use the query builder to investigate your environment. Every field in
BREAK DOWN draws a line, every
CALCULATE PER GROUP produces a separate graph.
LIMIT manipulate the results displayed.
Every Honeycomb account has a Fast Query Window that provides real-time querying for near term data. You choose what data, and how much of it, gets the high-performance treatment.
Store your data for as long as you like. Honeycomb’s customizable storage windows provide long term retention of events at a lower price point.
You can decide, dataset by dataset, how big to make your Fast Query Windows and how much Secondary Storage to provision.
Honeycomb gives you ways to dig ever deeper to zero in on an issue.
Interrogate the raw data behind query results and graphs. Pull events out of the noise with filters, like
endpoint_name starts_with “/product”. Sample incoming events to see the shape of the raw data. Create derived columns to graph more complex calculations of different dimensions and headers.
Bookmark and organize queries into shared boards. Use individual and team query histories to follow your own trail of breadcrumbs. Click a button to share a graph straight to Slack with your comment (and anyone who clicks through can play around with the same data).
Query results are stored forever, even after the data has aged out, so you can reliably pass them around or put them in docs and post-mortems.