Latency.at runs a set of probes around the world. Your Prometheus server can request performance and availability metrics about arbitrary targets from these probes. The requests are authenticated by a token which you can generate after signing up.
Let's say you want to get HTTPS metrics about the site example.com as seen from a probe running in Singapore at DigitalOcean. For that you would configure Prometheus to scrape:
At your defined scrape interval, Prometheus will connect to this URL. The prober handles the request by connecting to the specified target, using the specified module and return metrics about it. These metrics are now available in your server to be aggregated, correlated and graphed.
For this to work you need to set `bearer_token` to the auth token you generated in the dashboard. For more details, see the configuration section.
Depending on the target configuration, we return different metrics for HTTP, HTTPS, TCP and ICMP connections:
|*||probe_success||1 if request was successful|
|*||probe_duration||Duration of the request in seconds|
|*||probe_ip_protocol||IP protocol version used (4 or 6)|
|http||probe_http_duration_seconds||Duration of http request by phase, summed across all redirects (if any)|
|http||probe_http_status_code||Status Code of response|
|http||probe_http_redirects||Number of redirects followed during request|
|http||probe_content_length||Content length in response header|
|http(s)||probe_http_ssl||1 if HTTPS is used|
|https||probe_ssl_earliest_cert_expiry||Timestamp of earliest expiring certificate|
We will continuously extend the metrics we provide. If you're interested in specific protocols or metrics, let us know.
We operate probes in San Francisco, New York, Frankfurt, Bangalore and Singapore at DigitalOcean. For details see the configuration section. Soon we'll introduce more regions across multiple cloud providers. We plan to provide coverage in all regions served by the major cloud providers.
The `probe_duration` metric reflects the duration of the request, no matter whether it was successful or not. To calculate the duration of successful, or failed, requests only, use binary operators.
Calculate duration of successful requests:
probe_duration_seconds * (probe_success == 1)
Calculate duration of failed requests:
probe_duration_seconds + (probe_success == 0)
Our Exporter is based on the official Prometheus Blackbox Exporter. Beside operational metrics and integration in our accounting, we we provide more detailed HTTP metrics, most notable breakdown by request phases and redirect. We will try to contribute these changes back to the official Blackbox Exporter or release our open source exporter soon.
We're using Stripe to securly process your payments. After adding a credit card, you can subscribe to on of our plans.
On each renewal your request balance gets topped up by the given number of requests. Requests from the prior subscription cycle expire.
We are required to collect VAT from individual customer within the EU. If you're a business you can enter your VAT identification number in the dashboard's billing section to get exempt.
For existing subscription, the exemption get applies on the next billing cycle.
You can subscribe to a bigger plan any time and we will prorate the subscription costs and start a new billing cycle. This will add the number of request of the new plan to your account balance.
Example: You're on the $50 plan and decide to upgrade to the $100 plan in the middle of your billing cycle. We will credit you $50 for the unused portion of the billing cycle, charge you the different of $50 and start a new billing cycle. This will set your balance to 15,000,000. One month later, the subscription renews and you get charged the full $100.
You can also downgrade at any time, taking effect in the next billing cycle.
After downgrading you can still use up all your remaining credits until the end of your billing cycle. On the next cycle, you get charged the price of your new plan and unused credits expire.