kitebaseDeveloper platform
2 of 3night pages removed
Forge grouped a flood of flapping disk alerts into one capacity fix, and the on-call rotation went from broken sleep to quiet weeks.
Customers
Teams that sleep through the night.
From streaming to energy, engineering teams trust Forge with production. Here is what changes when an autonomous engineer joins the rotation.
- Fernvale
- Tidewell
- arcbright
- kitebase
- Solvane
- Mosshire
- Quillstone
- halcyra
What a quarter with Forge looks like.
- less time to resolve, on average
- 47%
- fewer pages reaching a person
- 8.6x
- lower cloud spend after one quarter
- 31%
- of drafted fixes merged by teams
- 96%
Median results across customers after their first 90 days with Forge.
The first week Forge found a cache regression we had been chasing for a month. Now it reads every deploy before we do, and our incident channel is mostly quiet.
61%
shorter incidents across 140 services
Median incident length
- Before Forge1h 54m
- After 90 days44m
We stopped writing runbooks for problems Forge already fixes. The engineers who used to firefight now spend that time on the product.
12 hrs
of on-call time returned each week
Pages reaching a person each week
- March 202664
- September 20269
More from the field.
- SolvaneEnergy analytics$410ksaved in the first year
Weekly Watches found idle clusters and oversized databases across four cloud accounts, each one shipped as a reviewed change.
- QuillstonePublishing11 minmedian time to root cause
Editors stopped noticing outages first. Forge traces regressions to the deploy that caused them before readers ever see an error.
Give production a steady hand.
Forge joins your team, learns how your environment really works and makes it a little more reliable every week.
