generic pool


generic object pool for node.

$ npm install apool

Initialize a new Pool with optional max.

Set max objects that the pool can hold.

Set the constructor that the pool will use when it needs more objects.

Set the destructor that will be used to destroy objects.

Set the min number of objects that will remain in the pool, defaulted to whatever is passed to populate().

Get the length() of objects (including used ones).

Populate n objects with optional fn(err).

Acquire an object with fn(err, obj) and optional timeout defaulting to 500ms.

Return an object to the pool.

  • add, emitted when an object is added.
  • return, emitted when an object is returned.
  • acquired, emitted when an object is acquired.
  • populate, emitted after populate(n).

See ./example, there are two real-world examples using phantomjs with a pool of page() objects, benchmarking each with wrk yields:


Running 30s test @ http://localhost:3000/
  12 threads and 100 connections
  Thread Stats   Avg      Stdev     Max   +/- Stdev
    Latency   457.23ms    9.24ms 459.79ms   97.25%
    Req/Sec     0.67      3.22    18.00     95.50%
  336 requests in 30.01s, 62.02KB read
  Socket errors: connect 0, read 0, write 0, timeout 1344
Requests/sec:     11.20
Transfer/sec:      2.07KB


Running 30s test @ http://localhost:3000/
  12 threads and 100 connections
  Thread Stats   Avg      Stdev     Max   +/- Stdev
    Latency   310.84ms   35.91ms 385.16ms   89.66%
    Req/Sec    25.23      4.51    38.00     80.32%
  9218 requests in 30.01s, 1.66MB read
Requests/sec:    307.16
Transfer/sec:     56.69KB
$ make test