We have a few platform-related questions regarding the calling APIs. We’re building automation for our clients, so we’d like to better understand how the calling infrastructure works in practice.
Call Latency: When we initiate a call through the API (either a single call or a batch), what is the typical latency before the call starts ringing on the recipient’s device? What factors affect this timing?
Concurrency Limits: How does call concurrency work? Is the maximum concurrency determined by the account or workspace plan? Does it vary between different account types?
Batch Processing: If our account has a maximum concurrency of 20 calls, and we initiate a batch of 20 calls, will all 20 calls begin dialing simultaneously, or are they processed sequentially with a delay between each call?
Multiple Batches: If the maximum concurrency is 20, can we start two separate batches of 20 calls at the same time from the same workspace? Or will the second batch remain queued until capacity becomes available?
Queue Behavior: If additional calls are queued because the concurrency limit has been reached, how does the platform decide when to start the next batch? Is there a fixed interval, or do queued calls begin automatically as soon as an active call slot becomes available?
These details will help us design our automation workflows more effectively and ensure we build the calling logic according to the platform’s best practices.
Hello @kavish Here’s a breakdown of each question based on the documentation:
1. Call Latency
CPS (Calls Per Second) is a key factor — it controls how quickly new calls are initiated. Your telephony provider also matters: Telnyx CPS changes take effect instantly, while Twilio changes take up to 10 minutes. The CPS limit itself paces how fast calls are dispatched.
2. Concurrency Limits
Concurrency limits apply per workspace, not per account. Pay-As-You-Go workspaces default to 20 concurrent calls. You can purchase additional concurrency or enable Burst Mode (up to 3× your limit or limit + 300, whichever is lower) for a $0.10/min surcharge on burst calls. Limits are adjustable from Settings > Limits.
3. Batch of 20 Calls — Simultaneous or Sequential?
Calls in a batch are initiated as concurrency slots become available, governed by your CPS setting. With a default CPS of 1 (Twilio/Telnyx), calls would be dispatched 1 per second — not all at once. You can increase CPS (up to 16 for Telnyx, 5 for Twilio, 150 for Custom Telephony) to speed up dispatch.
4. Two Batches of 20 Simultaneously
The concurrency limit is workspace-wide. If your limit is 20 and both batches are active, the combined 40 calls share that pool — calls beyond the 20-slot limit will be queued, not rejected.
5. Queue Behavior
Queued calls begin automatically as soon as a concurrency slot becomes available — there is no fixed interval. The batch status shows as “Ongoing” while calls are being initiated as slots open, and moves to “Sent” once all calls complete.