This image serves as a simple example how to create a Docker image for a NodeJS application and includes some endpoints to break the application through OOM kills or provoke slow requests.
| Path | Method | Description |
|---|---|---|
/ |
GET |
Returns some hello world JSON data. |
/random-nr |
GET |
Returns a random number (statically set to 4, so it can be fixed during training) |
/slow-request |
GET |
Purposefully slow request. Set the time it waits in seconds (rational numbers allowed) using the request_time parameter in the query string, e.g. /slow-request?request_time=2 |
/allocate-memory-kb |
POST |
Allocates a string of data and pushes it to a global array, increasing memory usage. This can be used to provoke deliberate OOM kills if memory is limited. Amount of memory must be specified in a JSON request body using the memory_size_kb property. E.g. curl -X POST -H "Content-Type: application/json" --data '{ "memory_size_kb": 1 }' localhost:8080/allocate-memory-kb allocates 1 KB. |
/healthcheck |
GET |
Healthcheck endpoint that returns an alive message and can be configured to leak some memory on each request (see. configuration). |
/metrics |
GET |
Returns metrics in ini style format for prometheus. |
| Env Variable | Values | Default | Description |
|---|---|---|---|
PORT |
Integer from port range | 8080 | Specifies the port where the server listens. |
HEALTHCHECK_LEAK_MEMORY |
{ empty | FALSE | "TRUE" } | empty => FALSE | Whether a request to the healthcheck endpoint should leak some memory. |
HEALTHCHECK_LEAK_INCREMENT_KB |
Integer | 1000 | Amount of memory that should be allocated on each healthcheck request if HEALTHCHECK_LEAK_MEMORY is "TRUE" in kilobytes. |