diff --git a/dependencies/stdout-log-generator/README.md b/dependencies/stdout-log-generator/README.md
index dd8559361a..9732c97630 100644
--- a/dependencies/stdout-log-generator/README.md
+++ b/dependencies/stdout-log-generator/README.md
@@ -1,19 +1,86 @@
 # Stdout Log Generator
 
-Small tool for generating logs to stdout, used in the e2e tests of the telemetry-manager.
+A small tool that continuously writes log lines to stdout. Used in the e2e tests of telemetry-manager to produce configurable log traffic inside Kubernetes workloads.
 
-Available command line args:
+The tool exposes a Prometheus metrics endpoint on port `2112` (`/metrics`) with two metrics:
 
-- format
-- bytes
-- rate
-- fields
-- text
+- `logs_generated_total` — total log lines written since start
+- `logs_generated_rate` — actual throughput in logs/second since start
 
-## Build locally
+## Usage
+
+```sh
+stdout-log-generator [flags]
+```
+
+## Flags
+
+| Flag | Short | Default | Description |
+|------|-------|---------|-------------|
+| `--format` | | `json` | Output format: `json` or `plaintext` |
+| `--bytes` | `-b` | `2048` | Size of each log line in bytes |
+| `--rate` | `-r` | `1` | Target log lines per second. `0` means unlimited |
+| `--fields` | `-f` | | Additional key=value fields to include in every JSON log record (can be repeated or comma-separated). Ignored when `--format plaintext` is used |
+| `--text` | `-t` | | Fixed text to write for every plaintext log line. Ignored if `--bytes` is set. Only relevant when `--format plaintext` is used |
+
+## Output formats
+
+### JSON (default)
+
+Each line is a JSON object with a `padding` field filled with random characters to reach the requested byte size, plus any custom fields added via `--fields`:
+
+```json
+{"padding":"aBcDeFgH...","environment":"prod","app":"myservice"}
+```
+
+The `padding` field is sized so that the total serialized JSON line matches `--bytes` exactly (including the trailing newline added by the JSON encoder).
+
+If the combined size of the custom fields already exceeds `--bytes`, the tool exits with an error.
+
+### Plaintext
+
+Each line is either the fixed string given by `--text`, or a random string of `--bytes` length when `--text` is not set.
+
+## Rate limiting
+
+The tool uses a token-bucket rate limiter (`golang.org/x/time/rate`). On Linux, the minimum achievable sleep granularity is approximately 2 ms. To compensate, the burst size is set to `2 × (rate / 1000)` so that the generator can emit multiple logs back-to-back before sleeping, keeping the long-run average close to the requested rate even at high throughputs.
+
+Setting `--rate 0` disables throttling entirely (write as fast as possible).
+
+## Examples
+
+Generate 100 JSON logs per second, each 1 KiB in size:
 
-To build the image locally run:
+```sh
+stdout-log-generator --rate 100 --bytes 1024
+```
+
+Generate JSON logs with custom fields at the default rate:
+
+```sh
+stdout-log-generator --fields app=myservice,env=prod
+```
+
+Generate plaintext logs with a fixed message at 50 logs/second:
+
+```sh
+stdout-log-generator --format plaintext --text "hello world" --rate 50
+```
+
+Generate random plaintext logs of 512 bytes each, unlimited rate:
+
+```sh
+stdout-log-generator --format plaintext --bytes 512 --rate 0
+```
+
+## Build locally
 
 ```sh
 docker build -t stdout-log-generator:local .
 ```
+
+Run the local image:
+
+```sh
+docker run --rm stdout-log-generator:local --rate 10 --fields app=test
+```