Fault library
Partitions, clock skew, disk pressure, bad DNS, half-open sockets.
01Chaos testing for stateful systems
INJECT THE FAILURE YOU ARE MOST AFRAID OF — IN STAGING, ON PURPOSE, WITH A KILL SWITCH.
Glitchcrafted — Chaos testing for stateful systems. No ceremony, no sales deck, no “book a call to see pricing”.
Overview
Runs
Policies
Traces
Settings
Chaoslive
1split-brain found on a Tuesday
0surprises during peak season
Fault libraryok
Blast radius controlok
Scheduled by riskok
Teams shipping with Glitchcrafted
NorthwindKestrelRuangPolarTokoLumenBakauVela
What you get
Three things Glitchcrafted does properly.
Blast radius control
Every experiment names its scope and refuses to start outside it.
02Scheduled by risk
Run the experiment that matches the incident you had last quarter.
03“We found the bug we had been dreading, on a quiet afternoon, with the whole team watching.”
Tomasz W. — SRE, Berlin
Inside the product
One screen, the whole picture.
Overview
Runs
Policies
Traces
Settings
Chaoslive
1split-brain found on a Tuesday
0surprises during peak season
Fault libraryok
Blast radius controlok
Scheduled by riskok
1split-brain found on a Tuesday
0surprises during peak season
23fault types included
Questions
The things people ask first.
Is it safe in production?+
Start in staging. Production runs require an explicit scope and a stop condition.
What if it breaks something?+
That is the point — but every experiment has a kill switch and a hard timeout.
Do you need agents installed?+
One lightweight daemon per host, or a Kubernetes operator.
Pricing
Start free. Pay when it earns it.
Glitchcrafted is open for design partners.
Small team, direct access to the people who build it. Bring a real workload.
Request access