In @components/api-server/src/routes.rs:
- Around line 60-64: The ReverseProxy instantiation hardcodes the hostname
"log_ingestor"; update it to read host from the configuration instead: add a
host field to the log_ingestor config (e.g., config.log_ingestor.host) in
clp_rust_utils, then use that host together with config.log_ingestor.port when
building the URL passed to axum_reverse_proxy::ReverseProxy::new (replace
"log_ingestor" with the configurable host using format! to construct
"http://{host}:{port}"). Ensure the config type and any deserialization/defaults
are updated so host is available where ReverseProxy::new is called.
- Around line 33-36: Update the doc comment in
components/api-server/src/routes.rs where the Panics section reads "failes" and
correct the typo to "fails" so the sentence reads "Panics if the log ingestor is
enabled but fails to install `aws_lc_rs` as the crypto provider."; ensure you
only change the misspelled word in the existing doc comment (the Panics block)
and leave the rest unchanged.
- Around line 57-59: The call to
rustls::crypto::aws_lc_rs::default_provider().install_default() currently uses
.expect(...), which will panic if a provider is already installed; change it to
handle the Err variant instead (e.g., check if let Err(_) =
rustls::crypto::aws_lc_rs::default_provider().install_default() and log a debug
or warn message like "Crypto provider already installed by another component")
so the API server doesn't crash on duplicate provider installation.

In @components/api-server/src/routes.rs:
- Around line 52-55: Add a concise comment above the early return that checks
LogsInput::S3 explaining why the reverse proxy is only required for S3 log
input: note that when config.logs_input is S3 the service must proxy requests to
the external S3-backed log ingestor (or internal ingestor accessing S3 objects)
and therefore the reverse proxy/router needs to be set up, whereas other
LogsInput variants do not require that HTTP proxying; reference the
LogsInput::S3 check and the router variable so future readers understand the
S3-specific dependency on the log ingestor and why we short-circuit for other
inputs.
- Around line 57-64: The code adds the aws_lc_rs crypto provider because
axum_reverse_proxy::ReverseProxy enables TLS by default even when the target URL
uses http://; either keep the existing
rustls::crypto::aws_lc_rs::default_provider().install_default() call or remove
that provider and disable TLS in the axum-reverse-proxy dependency to avoid
pulling the TLS lib unnecessarily; to change, update the Cargo.toml entry for
axum-reverse-proxy to explicitly set features = [] (e.g., axum-reverse-proxy = {
version = "1.1.1", features = [] }) or retain the install_default() call near
where ReverseProxy::new("/log_ingestor", &format!("http://log_ingestor:{port}"))
is created so the required crypto provider is present.

In @components/api-server/src/routes.rs:
- Around line 63-70: The code reads LOG_INGESTOR_HOSTNAME_ENV_VAR into hostname
but doesn't guard against an empty string, causing a malformed URL; update the
logic around hostname to treat empty or whitespace-only values as missing: fetch
the env var via std::env::var(LOG_INGESTOR_HOSTNAME_ENV_VAR), .map(|s|
s.trim().to_owned()), then if the result is Err or the trimmed string is empty
log the existing info message and set hostname =
DEFAULT_LOG_INGESTOR_HOSTNAME.to_owned(); ensure references to
LOG_INGESTOR_HOSTNAME_ENV_VAR, DEFAULT_LOG_INGESTOR_HOSTNAME and the hostname
variable are used so downstream URL construction uses a valid non-empty
hostname.