Telephony

Concurrency and limits

How many calls run at once, and what happens past that.

What concurrency means

Concurrency is the number of calls running at the same moment. A call takes a slot when it starts and frees it when it ends.

Concurrency is not calls per day. Five calls at once, each lasting two minutes, finish about 150 calls an hour.

What your plan includes

Plan	Concurrent calls
Pay as you go	20 included. Extra capacity is $8 per concurrent call per month.
Enterprise	Custom

Settings → Plan and billing shows your limit under Concurrent calls: Up to 20 on Pay as you go, Custom on Enterprise. To add capacity or move to Enterprise, click Contact sales there.

Two limits apply to outbound

An outbound deployment is bounded by your plan and by its numbers.

Limit	Value
Your plan	The concurrent calls above, across your workspace.
Per deployment	Max parallel calls, set when you deploy. It can't exceed 5 × the numbers you deploy with.

So a deployment on 4 numbers can run at most 20 calls at once, even if your plan allows more. To run more calls in one deployment, deploy with more numbers. See number-pools.

Deployments created through the API use one number and run 1 to 5 calls at once (max_parallel_calls). See api-deployments.

Estimate your peak

peak concurrent calls ≈ calls per hour × average call length in minutes ÷ 60

1,000 calls of 2 minutes spread over 4 hours is 250 calls an hour, so about 8 at once. An inbound line taking 30 calls an hour at lunch, 4 minutes each, peaks at about 2.

Measure real call lengths in monitoring and call-logs before you size for a campaign.

Concurrency is not the API rate limit

Concurrency counts calls in progress. API rate limits count HTTP requests per API key per minute, and each endpoint group has its own limit. A request over the limit returns 429. See api-rate-limits.

What goes wrong

Symptom	Cause
Max parallel calls won't go higher	It is capped at 5 per number in the deployment. Add numbers.
A large campaign takes longer than planned	Too few numbers in the deployment, or Max parallel calls set low.
`429` from the API	You hit an API rate limit, not a call limit. Wait and retry. See api-rate-limits.