refactor(python-sdk): unify the pyqwest connection pools

.changeset/python-unify-pyqwest-connection-pools.md

@@ -0,0 +1,17 @@
+---
+"@e2b/python-sdk": patch
+---
+
+Run every persistent HTTP stack in the SDK on one shared pyqwest connection pool
+instead of four: the control-plane REST API, the envd HTTP API, the envd RPC
+clients, and the volume content API now all draw from
+`e2b.api.client_sync`/`client_async`, keyed on the three knobs that are fixed
+when a pyqwest transport is built — proxy, idle read bound, and HTTP version.
+reqwest pools per host internally, so one pool serves the API host and every
+per-sandbox host without interference — and since envd RPC and the envd HTTP API
+hit the same host, an active sandbox now needs a single HTTP/2 connection instead
+of one per stack. Streamed downloads keep a pool of their own, the only one
+carrying the idle `read_timeout`: reqwest's read timer runs during body send and
+TTFB, so on a shared pool it would cut off long uploads. No signature changes —
+`get_transport` and `get_envd_transport` keep the `http2` parameter restored in
+2.39.1, and the two are now the same pool per key rather than two.

packages/python-sdk/e2b/api/client_async/__init__.py

@@ -24,146 +24,179 @@ def get_api_client(config: ConnectionConfig, **kwargs) -> AsyncApiClient:
 
 
 class ConnectionRetryTransport(RetryTransport):
-    """Retry only failures establishing the connection — shared by the REST
-    API and envd RPC stacks: pyqwest raises the builtin ``ConnectionError``
-    only before the request was written, so these retries can never replay a
-    request the server may have received (a delivered REST call or unary RPC
-    like ``SendInput``). This matches the connect-only ``retries`` of the
-    httpx transports this replaced; the retry middleware's default policy
-    would otherwise also retry I/O errors and 429/5xx responses for
-    idempotent methods."""
+    """Retry only failures establishing the connection — part of the shared
+    transport stack, so it covers the REST API, the envd HTTP API, the envd RPC
+    clients and the volume content API alike: pyqwest raises the builtin
+    ``ConnectionError`` only before the request was written, so these retries
+    can never replay a request the server may have received (a delivered REST
+    call or unary RPC like ``SendInput``). This matches the connect-only
+    ``retries`` of the httpx transports this replaced; the retry middleware's
+    default policy would otherwise also retry I/O errors and 429/5xx responses
+    for idempotent methods."""
 
     def should_retry_response(
         self, request: Request, response: Union[Response, Exception]
     ) -> bool:
         return isinstance(response, ConnectionError)
 
 
-def retrying_http_transport(
+_TransportKey = Tuple[Optional[ProxyConfig], Optional[float], bool]
+"""Cache key of the shared transports: proxy, idle read bound, HTTP version.
+
+All three are fixed when a pyqwest transport is constructed, so each distinct
+combination is necessarily its own pool."""
+
+_transport_lock = threading.Lock()
+# One pyqwest transport — one reqwest connection pool — per key; a `None` proxy
+# is the direct pool. Every HTTP stack in the SDK draws from these: the
+# control-plane REST API, the envd HTTP API, the envd RPC clients
+# (`e2b.envd.client_async`) and the volume content API
+# (`e2b.volume.client_async`). reqwest pools per host internally, so a single
+# pool serves the API host and every per-sandbox host without interference —
+# and because envd RPC and the envd HTTP API share it, a sandbox needs one
+# HTTP/2 connection instead of one per stack.
+#
+# pyqwest's I/O runs on its own Rust runtime, so unlike the httpx transports
+# they replaced, the transports are not bound to an event loop and the caches
+# are process-global rather than per-loop.
+_transports: Dict[_TransportKey, ConnectionRetryTransport] = {}
+# The httpx adapter over each pool, shared by every httpx client on it.
+_httpx_transports: Dict[_TransportKey, AsyncPyqwestTransport] = {}
+
+
+def get_pyqwest_transport(
     proxy: Optional[ProxyConfig],
     read_timeout: Optional[float] = None,
     http2: bool = True,
 ) -> ConnectionRetryTransport:
-    """A fresh pyqwest transport (= its own connection pool) with the SDK's
-    shared tuning — system CA certs (without which TLS through an
-    intercepting proxy fails), the httpx-equivalent pool limits, and
-    connect-only retries. The REST API, envd RPC, and envd HTTP API stacks
-    each cache their own instances (pool unification is a follow-up).
-
-    ``read_timeout`` bounds every read on the transport's connections; see
-    :func:`get_envd_transport` for when that is (and isn't) appropriate.
-
-    ``http2=False`` pins the transport to HTTP/1.1; see :func:`get_transport`
-    for when that matters.
+    """The shared pyqwest transport (= one connection pool) with the SDK's
+    tuning — system CA certs (without which TLS through an intercepting proxy
+    fails) and the httpx-equivalent pool limits — behind connect-only retries.
+
+    Consumers speaking pyqwest natively (the envd RPC clients) take this;
+    consumers speaking httpx take :func:`get_httpx_transport`, the adapter over
+    the very same pool. Layer concerns above it rather than into it — the RPC
+    stack's plain-HTTP-error normalization wraps it, headers and codecs are
+    per-request — so that the pool stays shareable.
+
+    ``read_timeout`` bounds every read on the pool's connections and ``http2``
+    fixes the HTTP version; both are part of the cache key because they are
+    transport-construction knobs, so one pool cannot serve two values of
+    either. reqwest's read timer keeps running while a request body is sent and
+    while waiting for the response head, so a pool carrying one would cut off
+    long uploads and slow responses — only streamed downloads ask for it, as an
+    idle bound (see :func:`get_transport`).
 
     Requests are logged by pyqwest itself on the ``pyqwest.access`` and
     ``pyqwest`` loggers at ``DEBUG`` (off unless enabled) — the transport-level
     diagnostics httpcore used to provide. The SDK's own ``logger`` option is
     separate and sits above this, on the httpx client."""
-    return ConnectionRetryTransport(
-        HTTPTransport(
-            tls_include_system_certs=True,
-            proxy=proxy.to_pyqwest() if proxy is not None else None,
-            pool_idle_timeout=pool_idle_timeout,
-            pool_max_idle_per_host=pool_max_idle_per_host,
-            read_timeout=read_timeout,
-            # `None` leaves the version to ALPN on TLS connections (HTTP/2
-            # against the E2B API) and uses HTTP/1 for plaintext, like the
-            # http2-enabled httpx transport this replaced.
-            http_version=None if http2 else HTTPVersion.HTTP1,
-            # Redirects belong to the httpx client above (which the generated
-            # clients leave off), not to reqwest.
-            follow_redirects=False,
-        ),
-        max_retries=connection_retries,
-    )
+    key = (proxy, read_timeout, http2)
+    with _transport_lock:
+        transport = _transports.get(key)
+        if transport is None:
+            transport = ConnectionRetryTransport(
+                HTTPTransport(
+                    tls_include_system_certs=True,
+                    proxy=proxy.to_pyqwest() if proxy is not None else None,
+                    pool_idle_timeout=pool_idle_timeout,
+                    pool_max_idle_per_host=pool_max_idle_per_host,
+                    read_timeout=read_timeout,
+                    # `None` leaves the version to ALPN on TLS connections
+                    # (HTTP/2 against the E2B API and envd) and uses HTTP/1 for
+                    # plaintext, like the http2-enabled httpx transport this
+                    # replaced.
+                    http_version=None if http2 else HTTPVersion.HTTP1,
+                    # Redirects belong to the httpx client above (which the
+                    # generated clients leave off), not to reqwest.
+                    follow_redirects=False,
+                ),
+                max_retries=connection_retries,
+            )
+            _transports[key] = transport
+        return transport
 
 
-_transport_lock = threading.Lock()
-# One transport (= one connection pool) per (proxy, http2) pair; a None proxy
-# is the direct pool. pyqwest's I/O runs on its own Rust runtime, so unlike
-# the httpx transports they replaced, the transports are not bound to an event
-# loop and the caches are process-global rather than per-loop.
-_transports: Dict[Tuple[Optional[ProxyConfig], bool], AsyncPyqwestTransport] = {}
+def get_httpx_transport(
+    proxy: Optional[ProxyConfig],
+    read_timeout: Optional[float] = None,
+    http2: bool = True,
+) -> AsyncPyqwestTransport:
+    """The httpx adapter over the shared pool of
+    :func:`get_pyqwest_transport`, for the generated httpx clients (control
+    plane, envd HTTP API, volume content). The adapter holds no state of its
+    own and does not close the pool, so closing an httpx client leaves the
+    pool intact for the other clients on it."""
+    key = (proxy, read_timeout, http2)
+    # Resolve the pool before taking the lock: it takes the same one.
+    pool = get_pyqwest_transport(proxy, read_timeout, http2)
+    with _transport_lock:
+        transport = _httpx_transports.get(key)
+        if transport is None:
+            transport = AsyncPyqwestTransport(pool)
+            _httpx_transports[key] = transport
+        return transport
 
 
 def get_transport(
-    config: ConnectionConfig, http2: bool = True
+    config: ConnectionConfig, http2: bool = True, *, for_streaming: bool = False
 ) -> AsyncPyqwestTransport:
-    """The shared pyqwest-backed httpx transport for REST API calls. For TLS
-    connections ALPN negotiates the HTTP version (HTTP/2 against the E2B
-    API), like the http2-enabled httpx transport this replaced.
+    """The shared httpx transport for the control-plane REST API and the envd
+    HTTP API (file transfers, health checks) — one pool serves both, keyed by
+    the connection's proxy. For TLS connections ALPN negotiates the HTTP
+    version (HTTP/2 against the E2B API), like the http2-enabled httpx
+    transport this replaced.
 
     ``http2=False`` returns a separate transport (its own pool) pinned to
     HTTP/1.1. That matters for a server that reacts to a client going away:
     HTTP/2 multiplexes requests over one connection, so abandoning a request
     only resets its stream and the server may never notice, while HTTP/1.1's
     one-connection-per-request closes the connection and the server observes
-    the disconnect."""
-    proxy = proxy_to_config(config.proxy)
-    key = (proxy, http2)
-    with _transport_lock:
-        transport = _transports.get(key)
-        if transport is None:
-            transport = AsyncPyqwestTransport(
-                retrying_http_transport(proxy, http2=http2)
-            )
-            _transports[key] = transport
-        return transport
-
-
-# One transport per (proxy, http2, streaming) triple, separate from the REST
-# API pools — envd traffic goes to per-sandbox hosts.
-_envd_transports: Dict[
-    Tuple[Optional[ProxyConfig], bool, bool], AsyncPyqwestTransport
-] = {}
+    the disconnect.
+
+    ``for_streaming`` selects the pool carrying ``READ_TIMEOUT``, the idle
+    bound on every read: it resets after each successful read, so it caps how
+    long a streamed download may stall without limiting total transfer time.
+    It is fixed per pool — the adapter's per-request timeouts are
+    whole-request deadlines rather than idle bounds — so only streamed
+    downloads take it, and they get their own pool
+    (see :func:`get_pyqwest_transport`).
+    """
+    return get_httpx_transport(
+        proxy_to_config(config.proxy),
+        READ_TIMEOUT if for_streaming else None,
+        http2,
+    )
 
 
 def get_envd_transport(
     config: ConnectionConfig, http2: bool = True, *, for_streaming: bool = False
 ) -> AsyncPyqwestTransport:
-    """The shared pyqwest-backed httpx transports for the envd HTTP API
-    (file transfers, health checks).
-
-    The streaming transport carries ``read_timeout``, the idle bound on
-    every read: it resets after each successful read, so it caps how long a
-    streamed download may stall without limiting total transfer time. It is
-    fixed per transport — the adapter's per-request timeouts are
-    whole-request deadlines. Only streamed downloads use it: reqwest's read
-    timer keeps running while a request body is sent and while waiting for
-    the response head, so on the regular transport it would cut off uploads
-    and slow unary responses longer than the idle bound (those stay bounded
-    by their whole-request deadlines instead).
-
-    ``http2=False`` pins the transport to HTTP/1.1 — see
+    """The envd HTTP API's transport, which is :func:`get_transport` — the two
+    now share one pool per key, since reqwest pools per host and envd RPC and
+    the envd HTTP API hit the same sandbox host anyway.
+
+    Kept only as a backward-compatible alias of :func:`get_transport` for any
+    external importer of the older public name; nothing inside the SDK calls it
+    (``e2b-code-interpreter`` imports :func:`get_transport` directly). Prefer
     :func:`get_transport`.
+
+    :deprecated: Use :func:`get_transport` instead; will be removed in the next
+        major version.
     """
-    proxy = proxy_to_config(config.proxy)
-    key = (proxy, http2, for_streaming)
-    with _transport_lock:
-        transport = _envd_transports.get(key)
-        if transport is None:
-            transport = AsyncPyqwestTransport(
-                retrying_http_transport(
-                    proxy,
-                    read_timeout=READ_TIMEOUT if for_streaming else None,
-                    http2=http2,
-                )
-            )
-            _envd_transports[key] = transport
-        return transport
+    return get_transport(config, http2, for_streaming=for_streaming)
 
 
 def get_envd_api(
     config: ConnectionConfig, base_url: str, *, for_streaming: bool = False
 ) -> httpx.AsyncClient:
     """An httpx client for a sandbox's envd HTTP API (file transfers, health
-    checks) on the shared pyqwest transports. The client itself is a cheap
-    stateless wrapper — one per consumer is fine — while the pooled transport
-    underneath is shared and loop-independent."""
+    checks) on the shared transports. The client itself is a cheap stateless
+    wrapper — one per consumer is fine — while the pooled transport underneath
+    is shared and loop-independent."""
     return httpx.AsyncClient(
         base_url=base_url,
-        transport=get_envd_transport(config, for_streaming=for_streaming),
+        transport=get_transport(config, for_streaming=for_streaming),
         headers=config.sandbox_headers,
         event_hooks=make_async_logging_event_hooks(config.logger),
     )