Frequently Asked Questions

General

What is CallMeter?

CallMeter is a SaaS platform for testing and monitoring VoIP/SIP infrastructure. It simulates real SIP endpoints that register with your SIP servers, place calls, exchange audio and video media, and collect over 150 real-time quality metrics per endpoint per second. CallMeter is used for stress testing (how many concurrent calls can your system handle), quality assessment (what is the call quality under load), and continuous monitoring (is your SIP infrastructure healthy right now).

How does CallMeter work?

CallMeter uses distributed workers to simulate SIP endpoints. When you run a test, the platform allocates endpoints across available workers. Each worker creates real SIP user agents that register with your SIP registrar, place or receive calls using standard SIP signaling (RFC 3261), and exchange RTP media (audio and/or video) for the configured duration. During the call, workers collect detailed quality metrics from both the send and receive directions and report them to the platform in real time. After the test, you can analyze the results through interactive dashboards, time-series charts, and aggregate statistics.

What SIP methods does CallMeter support?

CallMeter supports the core SIP methods: REGISTER, INVITE, ACK, BYE, CANCEL, OPTIONS, UPDATE, PRACK, INFO (for DTMF), and REFER. This covers the full call lifecycle from registration through call setup, media exchange, and teardown. Additional methods and advanced SIP features (100rel, session timers, re-INVITE) are on the product roadmap. See Roadmap for details.

Can I test private or internal SIP infrastructure?

Yes. Deploy a user-owned worker as a Docker container inside your private network. The worker connects outbound to the CallMeter platform (no inbound firewall rules required) and can reach your internal SIP servers, PBXs, and SBCs. This is the recommended approach for testing infrastructure that is not exposed to the public internet. See Deploying Your Own Workers.

Tests and Probes

What is the difference between a test and a probe?

A test is a one-time execution of a SIP load scenario. You define the number of endpoints, call duration, codecs, and groups, then run it when you want results. Tests are used for stress testing, capacity planning, and quality benchmarking.

A probe is a scheduled, repeating test that runs automatically at a defined interval (every 5, 15, 30, or 60 minutes). Probes are designed for continuous monitoring. They evaluate quality thresholds after each execution and report health status (HEALTHY, DEGRADED, UNHEALTHY, UNKNOWN). Probes can trigger webhooks on status changes and power public status pages. See Creating a Probe.

How many concurrent endpoints can I run?

The maximum number of concurrent endpoints depends on your plan. Plans range from small (suitable for basic testing and monitoring) to enterprise (thousands of concurrent endpoints). Check the Plans and Pricing page for current limits. If you need more capacity than the highest plan offers, contact the CallMeter team for a custom arrangement.

What is a group in a test?

A group is a subset of endpoints within a test that share the same configuration: registrar, codecs, media files, region, and worker assignment. Groups enable advanced testing scenarios:

• Cross-registrar testing: Group A registers with Registrar 1, Group B registers with Registrar 2. Group A calls Group B to test inter-registrar routing.
• Mixed codec testing: Group A uses Opus, Group B uses G.711, to compare quality under the same network conditions.
• Geographic distribution: Group A uses workers in US East, Group B uses workers in EU West, to test cross-region call quality.
• Bidirectional testing: Group A acts as callers, Group B acts as callees.

See Creating a Test for detailed configuration.

Can I test video calls?

Yes. CallMeter supports H.264 and VP8 video codecs. When creating a test, enable video on a per-group basis and configure the codec, bitrate, resolution, and frame rate. CallMeter collects video-specific metrics including frame rate, freeze events, keyframe requests (PLI, FIR), NACK counts, and decode errors. See Supported Codecs.

What is bidirectional testing?

Bidirectional testing means one group of endpoints acts as callers and another group acts as callees. The callers dial the callees using SIP, and both sides exchange media. This tests the complete call flow including SIP routing between the two groups. Without bidirectional testing, CallMeter endpoints call out to your SIP infrastructure but do not call each other.

Can I test DTMF and IVR systems?

Yes. CallMeter supports DTMF generation and detection via three transport methods: RFC 4733 RTP telephone-events (recommended), SIP INFO with application/dtmf-relay, or both simultaneously. Configure digit sequences (0-9, *, #) with per-digit timing control (50-1000ms).

For IVR testing, use the scenario action framework to build event-driven call flows. SEND_DTMF actions transmit digits at specific times or in response to events. DTMF_RECEIVED triggers fire when incoming tones are detected, with optional per-digit filtering. Chain them together to navigate multi-level IVR menus automatically. DTMF metrics (dtmf_digits_sent, dtmf_digits_received) are tracked per endpoint. See Scenario Actions for details.

What happens if I stop a running test?

When you stop a running test, the platform initiates a graceful shutdown. All endpoints in the INCALL phase advance to DISCONNECTING and begin their teardown sequence (BYE sent). Endpoints that have not yet started their call phase are cancelled and move to CLOSED. Metrics collected up to the stop point are retained and available for analysis. The test run transitions to COMPLETED with partial results.

Workers

What regions are cloud workers available in?

Cloud workers are deployed in multiple geographic regions. The exact regions available depend on current platform capacity. Check the Workers page in your project to see available regions when creating a test. For testing SIP infrastructure in a region without cloud workers, deploy a user-owned worker in that location.

How do I deploy my own worker?

User-owned workers run as Docker containers. The deployment process is:

1. Create a worker in the CallMeter platform (Workers page) and receive a cmw_ authentication token
2. Deploy the Docker container on your infrastructure with the token as an environment variable
3. The worker connects outbound to the CallMeter platform and appears as ONLINE on the Workers page

See Deploying Your Own Workers for complete instructions including Docker commands and configuration options.

What happens if a worker disconnects during a test?

If a worker loses its connection to the CallMeter platform during a test run, the platform detects the missing heartbeat within 60 seconds. Endpoints assigned to the disconnected worker are moved to CLOSED phase with an OSERROR outcome. If other workers are also executing endpoints for the same test, those endpoints continue normally and the test completes with partial results. If the disconnected worker was the only one in the run, the test transitions to FAILED status. See Worker Statuses.

Can I use both cloud and user-owned workers in the same test?

Yes. Different groups within the same test can use different workers. For example, Group A can use a cloud worker in US East while Group B uses a user-owned worker inside your data center. This enables testing call flows that span public and private infrastructure.

Metrics

What metrics does CallMeter collect?

CallMeter collects over 150 real-time metrics per endpoint per second, organized into seven categories:

• Quality: MOS, R-Factor, jitter, RTT, packet loss, clock drift
• Network: Packets sent/received, bitrate, duplicate packets, out-of-order packets, sequence gaps
• Feedback: NACK, PLI, FIR, REMB (RTCP feedback messages)
• Jitter Buffer: Buffer delay, target level, late packets, overflow/underflow events
• Audio: PLC events, audio level, DTMF detection, comfort noise
• Video: Frame rate, freeze events, decode errors, resolution changes
• Call Timing: PDD, setup time, ring time, call duration, ASR, NER

All metrics are recorded per direction (send and receive) and are available as time-series charts on the endpoint detail page. See Metrics Reference.

How is MOS calculated?

CallMeter calculates MOS using the E-model defined in ITU-T G.107. The calculation takes into account jitter, packet loss, round-trip time, and codec-specific impairment factors. The result is an estimated MOS on the 1 to 5 scale. This is a computed estimate, not a perceptual measurement (PESQ/POLQA). Perceptual scoring is on the product roadmap. See MOS Score.

Can I export test results?

Test results can be accessed programmatically via the CallMeter API. The API provides endpoints for retrieving test run summaries, endpoint lists, and metric data. API results are returned in JSON format and can be integrated into your own reporting, dashboards, or data analysis tools. See API Endpoints.

Billing and Plans

How are minutes counted for billing?

CallMeter bills based on stream-minutes. Each active audio or video stream consumes minutes independently. A single endpoint with audio only uses one audio stream-minute per minute of call duration. An endpoint with audio and video uses one audio stream-minute plus one video stream-minute per minute. Billing is calculated per second and rounded up. See Plans and Pricing and Credits and Overages.

What happens if I exceed my plan limits?

When you exceed your plan's included minutes, additional usage is billed as overage at your plan's per-minute rate. You can configure spending limits and alerts to prevent unexpected charges. Usage is tracked in real time and visible on the billing dashboard. See Credits and Overages.

Can I change my plan?

Yes. You can upgrade or downgrade your plan at any time from the billing settings page. Upgrades take effect immediately. Downgrades take effect at the end of the current billing period. See Managing Subscription.

Integrations

Can I integrate CallMeter with CI/CD pipelines?

Yes. CallMeter provides a REST API that can be called from CI/CD pipelines. You can create tests, trigger runs, poll for completion, and check quality thresholds programmatically. This enables automated quality gates in your deployment pipeline: if a post-deployment SIP test shows degraded MOS or high packet loss, the pipeline can fail the deployment. See CI/CD Integration.

How do I set up webhook notifications?

Webhooks are configured per probe. When a probe's health status changes (e.g., HEALTHY to DEGRADED), CallMeter sends an HTTP POST to your configured webhook URL with a JSON payload containing the probe ID, previous status, new status, timestamp, and metric values.

Webhook payloads are signed with HMAC-SHA256 for security verification. You can integrate webhooks with Slack, PagerDuty, OpsGenie, or any system that accepts HTTP POST requests. See Webhooks.

What is a status page?

A status page is a public-facing web page that displays the current health of your SIP infrastructure based on probe results. You can create status pages that show one or more probes, each with its current health state (Operational, Degraded, Major Outage). Status pages are accessible via a public URL and can be shared with customers, partners, or stakeholders. See Status Pages.

Does CallMeter have a CLI?

CallMeter offers API-based integration that can be used with standard CLI tools like curl. Combined with CI/CD integration patterns, you can automate test execution from the command line. See API Authentication and API Endpoints.

SIP Accounts and Registrars

How does SIP account pooling work?

When you configure a registrar with multiple SIP accounts and run a test with more endpoints than accounts, CallMeter uses account pooling. Each SIP account is assigned to one endpoint at a time. If there are more endpoints than accounts, the platform assigns accounts to as many endpoints as possible and notifies you of the shortfall.

For optimal testing, configure at least as many SIP accounts as the maximum number of concurrent endpoints in any single group that uses that registrar. See SIP Accounts.

Can I use CallMeter's built-in registrar?

CallMeter provides a cloud registrar with ephemeral SIP accounts for testing scenarios where you do not need to test against your own infrastructure. This is useful for:

• Initial platform exploration and learning
• Testing worker connectivity and basic call flow
• Bidirectional tests where both sides use CallMeter endpoints

For production testing of your own SIP infrastructure, you should configure your own registrars.

Security and Data

Is my data secure?

CallMeter implements multiple layers of security:

• Transport encryption: All connections between workers and the platform use encrypted channels
• Authentication: API keys, worker tokens, and user sessions are all independently authenticated
• Multi-tenant isolation: Each organization's data (tests, metrics, SIP traces) is strictly isolated. No cross-organization data access is possible.
• Credential storage: SIP account passwords are encrypted at rest
• Webhook security: Webhook payloads are signed with HMAC-SHA256 to prevent spoofing

Can I share test results with others?

Test results are accessible to all members of your organization who have the appropriate role. You can invite team members with different permission levels (Viewer, Editor, Admin). Viewers can see test results and metrics but cannot modify configurations. See Roles and Permissions.

Status pages provide a way to share probe health information publicly without requiring authentication. See Status Pages.

Where is my data stored?

CallMeter stores test configurations, SIP account credentials, and organizational data in its primary database. Time-series metrics (the 150+ measurements per endpoint per second) are stored in a purpose-built analytics database optimized for high-volume time-series data. All data is stored in secure, encrypted infrastructure.

How long is test data retained?

Data retention depends on your plan. Test run results, metrics, and SIP traces are retained for the period specified in your plan. After the retention period, data is automatically purged. Contact the CallMeter team for custom retention requirements.

Platform

What browsers are supported?

CallMeter's web interface supports the latest versions of Chrome, Firefox, Safari, and Edge. A modern browser with JavaScript enabled is required.

Can multiple team members use CallMeter simultaneously?

Yes. CallMeter supports multi-user organizations with role-based access control. Multiple team members can view test results, create tests, and manage probes concurrently. See Managing Members.

Is there a limit on the number of tests or probes I can create?

The number of tests and probes you can create depends on your plan. All plans allow multiple tests and probes. Check your plan details for specific limits on concurrent probes and saved test configurations. See Plans and Pricing.

How long does it take for test results to appear?

Metrics are collected in real time during the test and are available immediately on the test run detail page while the test is still running. After the test completes, aggregate statistics (averages, percentiles, MOS summaries) are computed within a few seconds. The full time-series data and endpoint-level detail are available for analysis as soon as the run transitions to COMPLETED.

What is the difference between a test run and a test configuration?

A test configuration is the saved definition of a test: endpoints, groups, codecs, registrars, duration. A test run is a single execution of that configuration. You can run the same configuration multiple times, and each run produces its own results. This lets you track quality changes over time by comparing runs of the same test.

Codec Selection

Which audio codec should I use for testing?

Use the codec that your production SIP infrastructure uses. If you are testing a PBX that handles G.711 (PCMA/PCMU) traffic, configure your test with G.711. If your infrastructure supports Opus, use Opus for higher quality and better packet loss resilience. For stress testing where you want to maximize the number of concurrent calls on a worker, G.711 uses less CPU than Opus because it requires no complex encoding. See Supported Codecs for the full list.

Which video codec should I use?

H.264 is the most widely supported video codec in SIP/VoIP infrastructure. Use H.264 if you are testing interoperability with existing systems. VP8 offers built-in error resilience and does not require patent licensing, making it a good choice for WebRTC-facing tests. If your SBC or gateway only supports one codec, match your test to that codec. Mismatched video codecs between caller and callee cause NEGOTIATION_FAILED outcomes.

Does the codec choice affect MOS scores?

Yes. The E-model MOS calculation includes a codec-specific impairment factor. G.711 (PCMA/PCMU) has a lower impairment factor (higher maximum MOS of approximately 4.4) than compressed codecs like G.729 (maximum MOS of approximately 3.9). When comparing MOS across tests, ensure both tests use the same codec, otherwise the comparison is not meaningful.

Billing Deep Dive

Do I get charged for failed tests?

You are billed for the actual stream-minutes consumed. If a test fails after 30 seconds, you are only billed for the 30 seconds of active media streams, not the full configured duration. Endpoints that fail to register and never establish media do not consume stream-minutes. See Credits and Overages.

Are probe executions billed the same way as tests?

Yes. Each probe execution is a small test run, and the stream-minutes consumed are billed at the same rate. A probe with 2 endpoints running every 15 minutes with 30-second calls uses a predictable amount of monthly minutes. You can estimate your monthly probe cost from the number of probes, their endpoint count, call duration, and interval.

What happens if I run out of credits mid-test?

If your credit balance reaches zero during a running test, the test completes normally. You will not be cut off mid-call. The overage is recorded and billed according to your plan's overage rate. To avoid unexpected overages, configure a spending limit on the billing settings page.

Workers Deep Dive

How many endpoints can a single worker handle?

It depends on the media configuration. For audio-only calls with G.711, a worker on a modern 4-core machine can handle several hundred concurrent endpoints. Adding video significantly increases CPU and memory requirements, reducing the per-worker capacity. Start with a small test and increase the endpoint count incrementally to find your worker's capacity limit. Monitor CPU and memory on the Docker host during tests.

Can I run a worker on a Raspberry Pi or ARM device?

CallMeter worker Docker images are available for x86_64 architecture. ARM-based devices are not currently supported. If ARM support is important to your use case, contact the CallMeter team to discuss it.

Do user-owned workers need a public IP address?

The worker itself does not need a public IP to connect to the CallMeter platform, it connects outbound. However, for SIP and RTP traffic, the worker needs network access to the SIP registrar and media endpoints. If those are on the public internet, the worker needs a routable path. If both the worker and the SIP infrastructure are on the same private network, no public IP is needed.

Security and Privacy Deep Dive

Does CallMeter see my SIP account passwords?

SIP account passwords are transmitted over encrypted connections and stored encrypted at rest. CallMeter workers use the credentials to authenticate with your SIP registrar during test execution. The credentials are never logged in plaintext and are not visible to CallMeter staff. You can rotate passwords at any time by updating the SIP account in CallMeter.

Can I restrict which team members can run tests?

Yes. CallMeter's role-based access control lets you assign different permission levels within your organization. Viewers can see results but cannot create or run tests. Editors can create and run tests. Admins can manage team members and billing. See Roles and Permissions.

Is SIP signaling data (call traces) stored?

SIP message traces from test runs are stored for the duration of your plan's data retention period. These traces contain SIP headers (including To, From, Contact, and Via) which may include SIP URIs and IP addresses. After the retention period, traces are automatically purged.

API Usage

How do I authenticate API requests?

API authentication uses API keys generated from your organization settings page. Include the API key in the Authorization header of each request. API keys are scoped to your organization and respect the same role-based permissions as the web interface. See API Authentication.

Can I trigger a test run from the API and wait for results?

Yes. The typical API workflow is: create or reference an existing test configuration, trigger a run, poll the run status endpoint until the run completes, then fetch the results. For CI/CD pipelines, you can set a timeout and check quality thresholds against the results to make pass/fail decisions. See CI/CD Integration.

Are there rate limits on the API?

Yes. The API enforces rate limits to protect the platform. The exact limits depend on your plan tier. If you exceed the rate limit, the API returns a 429 Too Many Requests response with a Retry-After header. Design your integrations to handle 429 responses gracefully with exponential backoff.

Related Pages

• Key Concepts -- Platform building blocks
• Quick Start -- Get started in minutes
• Roadmap -- Upcoming features
• API Endpoints -- Programmatic access