An external dependency observation product. It probes the third-party services your software depends on, from infrastructure you do not operate, records what each probe saw, and keeps a verifiable record of what happened.

How is an incident confirmed?

Deterministically, by persistence: 2 consecutive failed checks from the reliastra observation point open an incident, and 2 consecutive successes resolve it. One dropped probe is recorded, not declared.

How many places does it probe from?

One today. There is no regional quorum claim, because a quorum needs genuinely independent observation points and this deployment has one. When that changes, the site and the records will say so.

What is in an evidence record?

The dependency, the incident window and the observation topology behind it, every observation in the window, the arithmetic behind the availability figures, the attribution result with its methodology version, and a SHA-256 checksum over the payload.

Infrastructure you can prove.

RELIASTRA independently observes the external services your software depends on — and keeps a verifiable record of what they did.

Service starting at

$9/mo

Start monitoring Open public observatory

1 observation point today (us-east) · every record says so. No quorum claimed.

How evidence is produced

Your infrastructure includes services you do not operate.

How it works

Three steps, all published.

Observe

Every endpoint you register is checked every 300 seconds from infrastructure that is neither yours nor the vendor's. us-east is the observation point today, and every record says so.

Monitoring docs

Confirm

Failed checks are replayed against your incident window. 2 consecutive failures open an incident, 2 consecutive successes close it — no vendor status page involved.

Incident model

Prove

A resolved incident becomes a signed record: the window, every observation inside it, the availability arithmetic, and a SHA-256 checksum. Retained 365 days, verifiable by anyone.

Inside a record

Start monitoring Read the methodology

Public dependency index

The same probes, published.

What RELIASTRA's own probes observe. Not vendor-reported, no account required, every figure printed with its window and its observation count.

Open the observatory

Per vendor: current state, availability with its observation count, latency, incident history and the methodology that produced each figure.

Pricing

One plan. $9 a month.

Every capability the product has, on every account. No plan ladder, no seats, no annual negotiation, and no tier where the evidence gets better.

Start monitoring Terms and currency

The plan

$9/ month · USD

14-day trial on every new account. No payment method required to start, and nothing that expires into a locked account.

Dependencies: 25 monitored
Check interval: 30-second checks
Retention: 3 months

Product price: $9.00 (USD)
Actual charge: $9.00 (USD) per month
Payment provider: Paystack

RELIASTRA's plans are priced in USD. Our current Paystack payment flow processes payments in NGN. We are awaiting confirmation of additional payment options for international customers.

Questions, answered

The questions, answered the way the system answers them.

Full glossary →

What is RELIASTRA?: An external dependency observation product. It probes the third-party services your software depends on, from infrastructure you do not operate, records what each probe saw, and keeps a verifiable record of what happened.
How is an incident confirmed?: Deterministically, by persistence: 2 consecutive failed checks from the reliastra observation point open an incident, and 2 consecutive successes resolve it. One dropped probe is recorded, not declared.
How many places does it probe from?: One today. There is no regional quorum claim, because a quorum needs genuinely independent observation points and this deployment has one. When that changes, the site and the records will say so.
What is in an evidence record?: The dependency, the incident window and the observation topology behind it, every observation in the window, the arithmetic behind the availability figures, the attribution result with its methodology version, and a SHA-256 checksum over the payload.

Founder & principal engineer

Adeshina Emmanuel

AI and Cloud Infrastructure Security Engineer

Full profile

The engineer who operates the measurement network.

Adeshina Emmanuel is an infrastructure security engineer working across cloud identity, Kubernetes, AI systems, and the measurement systems that verify them. He architects and builds the RELIASTRA platform - the probe network, the Go CLI, the evidence pipeline - and publishes the method alongside the product.

The stack he designs and runs: Go for the CLI and probe tooling, Python and FastAPI for the API, Postgres and Redis, Celery for the scheduled measurement work, SHA-256 and Ed25519 for the evidence records.

Begin

Add one dependency you already own.

Read what the probe records for a day before deciding whether the rest of it is worth your attention. The trial needs no card.

Start monitoring See the public data first