Forecast CI feedback loop economics

What it's for

Shows CTOs what slow or unreliable CI is likely to cost—not only in runner spend, but in waiting time, failed releases, escaped defects and severe common-mode downside—using a model that must first work on their own later data.

What you give it

Inputs split into evidence read from your connected systems, calibration your team owns, and numerical controls that affect precision but never the result's meaning.

{
  "common_scenarios": [
    {
      "arrival_multiplier": 1,
      "common_escape_probability": 0,
      "common_failure_probability": 0,
      "id": "normal",
      "probability": 1,
      "runner_availability": 1,
      "service_time_multiplier": 1
    }
  ],
  "current_workloads": [
    {
      "compute_cost_per_minute": 0.1,
      "evidence_verified": true,
      "expected_run_count": 40,
      "id": "checkout-ci",
      "pipeline_class_id": "fast",
      "post_release_escape_loss": 10000,
      "runner_capacity_minutes": 1000,
      "terminal_failure_loss": 500,
      "value_per_feedback_minute": 2
    }
  ],
  "historical_periods": [
    {
      "clean_first_pass_count": 18,
      "compute_minutes_sum": 60,
      "eligible_run_count": 20,
      "evidence_verified": true,
      "feedback_log_minutes_sq_sum": 51.80580787960469,
      "feedback_log_minutes_sum": 32.18875824868201,
      "feedback_minutes_sum": 100,
      "id": "fast-0",
      "outcome_mature": true,
      "period_index": 0,
      "pipeline_class_id": "fast",
      "post_release_escape_count": 1,
      "recovered_flake_count": 1,
      "terminal_failure_count": 1,
      "timeout_or_cancel_count": 0
    },
    {

What you get back

This is the actual output of running the example above — computed by the same function the platform calls, not an illustration.

{
  "decision": "accepted",
  "detail_truncated": false,
  "interpretation": "The forecast is company-local and aggregate. Post-release escapes require governed incident linkage; association with a CI profile is not causal proof that a test or team caused an incident.",
  "method": "hierarchical_dirichlet_lognormal_ci_queue_economic_forecast",
  "portfolio_forecast": {
    "conditional_value_at_risk": 66309.5687,
    "expected_compute_cost": 13.1772,
    "expected_feedback_delay_cost": 437.2606,
    "expected_post_release_escape_loss": 24166.6667,
    "expected_terminal_failure_loss": 1428.3333,
    "expected_total_economic_loss": 26045.4378,
    "posterior_draws": 300,
    "tail_probability": 0.1,
    "value_at_risk": 51911.6219
  },
  "validation": {
    "accepted": true,
    "all_evidence_verified_and_mature": true,
    "baseline_feedback_mae_minutes": 7.5,
    "baseline_multiclass_log_loss": 0.96,
    "failure_ece": 0.0167,
    "holdout_periods": [
      5
    ],
    "holdout_row_count": 2,
    "holdout_run_count": 40,
    "maximum_failure_ece": 0.2,
    "minimum_validation_improvement": 0,
    "model_feedback_mae_minutes": 0.5556,
    "model_multiclass_log_loss": 0.7972,
    "relative_feedback_mae_improvement": 0.9259,
    "relative_log_loss_improvement": 0.1696,
    "training_period_count": 5,
    "training_run_count": 200,
    "unseen_current_class_ids": [],
    "unseen_holdout_class_ids": [],
    "unsupported_holdout_row_count": 0
  },
  "workloads": [
    {
      "expected_economic_loss": 26045.4378,
      "expected_post_release_escapes": 2.4167,
      "expected_terminal_failures": 2.8567,

How it works

Forecasting & survival — Estimate when something completes or fails, with censoring and unresolved work handled honestly rather than dropped.

1. 1 Construct consecutive zero-inclusive pipeline-class periods with mutually exclusive first-pass, recovered-flake, terminal-failure and timeout/cancel outcomes, mature post-release escape linkage, feedback log moments, compute and immutable evidence; reserve the latest whole periods before fitting.
2. 2 Partially pool class outcome probabilities, escape rates, feedback and compute toward company-wide priors, then require later-period multiclass log loss and feedback MAE to beat global baselines while failure calibration, class support, maturity and evidence gates pass.
3. 3 Draw one coherent arrival, service, runner-availability, common-failure and common-escape scenario per portfolio path; propagate queue amplification and price feedback, compute, terminal failures and mature escapes into expected loss, VaR and CVaR.

Before you trust it

Every tool in the catalog ships with the conditions under which its answer is meaningful — and the conditions under which it should abstain instead of guessing.

Calibrating it to your org

Calibration workflow

1. 1 Define the management decision, target outcome, aggregate unit, privacy boundary, cadence, and prediction/intervention horizon for this organization.
2. 2 Build a tenant-scoped historical cohort using only information available before each prediction or decision; preserve zero periods, censoring, assignment probabilities, and unresolved outcomes when the method requires them.
3. 3 Estimate statistical parameters on training history, but obtain costs, utilities, risk tolerance, practical-effect thresholds, capacity, and policy constraints from accountable decision owners.
4. 4 Validate on later time windows or held-out aggregate units at the deployment grain, against a simple baseline and the function-specific validation strategy.
5. 5 Deploy only if the returned decision clears evidence, overlap, calibration, robustness, and guardrail checks; warning, unsupported, schema-gap, and fallback decisions are abstentions.
6. 6 Monitor realized outcomes, data drift, coverage, and decision regret; recalibrate at a governed cadence or after a detected regime/definition change, never merely because a stakeholder dislikes the result.

Call it from your AI

You don't wire up 388 tools in your MCP client. The GitRevio MCP server exposes 18 tools, three of which let an agent search the catalog, read a tool's schema, and run it — so the assistant finds this one on its own.

gitrevio_capabilities_search
  { "q": "forecast companylocal ci feedback delay compute" }
  → finds "forecast_ci_feedback_loop_economics"

gitrevio_capability_describe
  { "capability_id": "forecast_ci_feedback_loop_economics" }
  → returns the input schema and agent guidance shown on this page

gitrevio_capability_run
  { "capability_id": "forecast_ci_feedback_loop_economics", "arguments": { ... } }
  → returns the result shown above

Works in Claude Desktop, Claude Code, Cursor, Cline, Continue.dev, Goose and Aider. See the MCP server.

Related tools

See every tool in Quality, incidents & reliability →