:hedgehog: PostHog is the leading platform for building self-driving products. Our developer tools – AI observability, analytics, session replay, flags, experiments, error tracking, logs, and more – capture all the context agents need to diagnose problems, uncover opportunities, and ship fixes. Steer it all from Slack, web, desktop, or the MCP.
Python ab-testingai-analyticsanalyticscdpdata-warehouseexperimentsfeature-flagsjavascriptproduct-analyticspythonreactsession-replaysurveystypescriptweb-analytics
import {
MessageFactory,
Rng,
TEAMS,
baseScenario,
finish,
group,
leaf,
pathRule,
setGeneratorClock,
severityRule,
} from '../parity/parity-scenarios'
import type { ParityMessage, ParityScenario } from '../parity/parity-types'
export type BenchWorkloadOptions = {
nowMs: number
batches: number
messagesPerBatch: number
recordsPerMessage: number
}
export const DEFAULT_BENCH_OPTIONS: Omit<BenchWorkloadOptions, 'nowMs'> = {
batches: 5,
messagesPerBatch: 200,
recordsPerMessage: 20,
}
// Buckets this large never bind, so a batch replayed many times does the same work every time.
const UNBOUNDED_LIMITER = {
LOGS_LIMITER_BUCKET_SIZE_KB: 10_000_000,
LOGS_LIMITER_REFILL_RATE_KB_PER_SECOND: 10_000_000,
}
const [plain, , , both, attr] = TEAMS
async function batchesOf(
rng: Rng,
options: BenchWorkloadOptions,
pickTeam: () => (typeof TEAMS)[number]
): Promise<ParityMessage[][]> {
const f = new MessageFactory(rng)
const batches: ParityMessage[][] = []
for (let b = 0; b < options.batches; b++) {
const batch: ParityMessage[] = []
for (let m = 0; m < options.messagesPerBatch; m++) {
batch.push(await f.logs(pickTeam(), options.recordsPerMessage))
}
batches.push(batch)
}
return batches
}
export async function buildBenchWorkloads(options: BenchWorkloadOptions): Promise<ParityScenario[]> {
setGeneratorClock(options.nowMs)
const workloads: ParityScenario[] = []
{
const s = baseScenario(
'passthrough',
'No settings or rules: the consumer forwards buffers undecoded.',
UNBOUNDED_LIMITER
)
s.batches = await batchesOf(new Rng(101), options, () => plain)
workloads.push(
finish(
s,
s.batches.map(() => options.nowMs)
)
)
}
{
const s = baseScenario('normalize', 'JSON body parse and PII scrub: decode, enrich, encode.', UNBOUNDED_LIMITER)
s.batches = await batchesOf(new Rng(102), options, () => both)
workloads.push(
finish(
s,
s.batches.map(() => options.nowMs)
)
)
}
{
const s = baseScenario('full', 'Every feature on for one team.', {
...UNBOUNDED_LIMITER,
LOGS_SAMPLING_ENABLED_TEAMS: '*',
LOGS_RETENTION_ENABLED_TEAMS: '*',
LOGS_PATTERN_MASKING_ENABLED_TEAMS: '*',
LOGS_METRICS_RULES_ENABLED_TEAMS: '*',
LOGS_DEFAULT_METRIC_RULES_ENABLED_TEAMS: '*',
LOGS_JSON_ATTRIBUTE_PARSING_ENABLED_TEAMS: '*',
LOGS_JSON_ATTRIBUTE_EXTRACTION_ENABLED_TEAMS: '*',
LOGS_BILLING_PRORATE_ENABLED: true,
})
s.teams = [{ ...attr, logs_settings: { ...attr.logs_settings, pii_scrub_logs: true } }]
s.sampling_rules = {
[attr.id]: [
severityRule('sev', { DEBUG: { type: 'drop' }, INFO: { type: 'sample', rate: 0.8 } }),
pathRule('path', { patterns: ['/healthz'], match_attribute_key: 'url.path' }),
],
}
s.retention_rules = {
[attr.id]: {
logs: [
{
id: 'errors',
config: {
retention_days: 90,
filter_group: group('AND', leaf('severity_text', 'exact', 'ERROR', 'log')),
},
},
],
},
}
s.metric_rules = {
[attr.id]: [
{
id: 'by-level',
metric_name: 'by_level',
filter_group: null,
value_attribute: 'count',
group_by: ['severity_text', 'service_name'],
},
],
}
s.batches = await batchesOf(new Rng(103), options, () => attr)
workloads.push(
finish(
s,
s.batches.map(() => options.nowMs)
)
)
}
{
const s = baseScenario('mixed', 'A spread of teams and settings, like a production batch.', {
...UNBOUNDED_LIMITER,
LOGS_SAMPLING_ENABLED_TEAMS: '102',
LOGS_PATTERN_MASKING_ENABLED_TEAMS: '101,104',
LOGS_RETENTION_ENABLED_TEAMS: '104',
})
s.sampling_rules = { 102: [severityRule('sev', { DEBUG: { type: 'drop' } })] }
const rng = new Rng(104)
s.batches = await batchesOf(rng, options, () => rng.pick(TEAMS))
workloads.push(
finish(
s,
s.batches.map(() => options.nowMs)
)
)
}
return workloads
}