Invariant

An invariant describes a state of the simulation log that must always hold. For example, here's one that expects no error-level signals:

type: sql
query: >
  select count(*) from signals
  where level = 'error';
expect: result == 0

Currently, the only supported type is "sql", with which you must specify:

• query - a SQLite-compatible query that pulls a value from the simulation log
• expect - a CEL expression that evaluates whether the invariant holds, based upon the result of the query

The query is run against the following schema:

CREATE TABLE IF NOT EXISTS effects (
    id INTEGER NOT NULL,
    key TEXT NOT NULL,
    offset INTEGER NOT NULL,
    value TEXT NOT NULL,
    format TEXT
);

CREATE TABLE IF NOT EXISTS signals (
    effect_id INTEGER,
    timestamp TEXT NOT NULL,
    system TEXT NOT NULL,
    level TEXT NOT NULL,
    data TEXT NOT NULL
);

CREATE TABLE IF NOT EXISTS errors (
    message TEXT NOT NULL
);

The CLI will evaluate the invariants while auditing the simulation once it completes. It will return an exit code of 1 and write something like this to stderr if any fail:

Audit
18/19 invariants passed
no-500s failed - got 3, expected 'result == 0'

Additionally, the CLI will write the results as json to .rngo/runs/last/invariants.json, e.g.:

{
  "no-500s": {
    "value": 3,
    "passed": false
  },
  "no-error-logs": {
    "value": 0,
    "passed": true
  }
}