Annotations¶
wxingress reads the nginx.ingress.kubernetes.io/* annotations of ingress-nginx with the semantics of ingress-nginx,
so that an Ingress written for ingress-nginx behaves the same without changes. It also keeps the nginx.org/*
annotations of NGINX Ingress Controller, on which it is built. This page lists which ingress-nginx annotations are
supported, which behave differently, and which are not supported. Names below are given without the
nginx.ingress.kubernetes.io/ prefix.
General rules¶
- An invalid value is ignored, with a Warning event on the Ingress, as ingress-nginx ignores it. Values are trimmed line by line first.
- Annotations that take nginx variables take only those every server knows:
$1to$9(the captures of the path's regular expression), the request's$http_<header>,$arg_<name>and$cookie_<name>, and$args,$binary_remote_addr,$content_length,$content_type,$document_root,$document_uri,$escaped_request_uri,$fastcgi_path_info,$fastcgi_script_name,$host,$hostname,$https,$is_args,$proxy_add_x_forwarded_for,$proxy_host,$proxy_port,$query_string,$realpath_root,$remote_addr,$remote_port,$remote_user,$request_filename,$request_id,$request_method,$request_uri,$scheme,$server_addr,$server_name,$server_port,$server_protocoland$uri. Any other variable (also$best_http_host, which wxingress does not define) makes the value invalid, since nginx would refuse the whole configuration. Regular expressions nginx compiles (paths withuse-regexorrewrite-target,~values ofproxy-redirect-fromandproxy-cookie-*) must be valid PCRE; .NET-only syntax such as balancing groups is refused. - Where an
nginx.org/*annotation sets the same thing, thenginx.org/*annotation wins and the ingress-nginx one is ignored with a warning (see nginx.org annotations). nginx.org/*,nginx.com/*andappprotect.f5.com/*annotations are validated differently: most of them are checked before the Ingress is configured, and an invalid value - or one that needs NGINX Plus, App Protect or-enable-snippets- rejects the whole Ingress (Warning eventRejected) instead of being ignored.- Where ingress-nginx refuses to serve a location because its protection cannot be set up (a missing auth Secret, an
invalid
auth-url, an invalid source range, a missing ConfigMap), the locations of the Ingress answer 503 instead of serving without it. The Warning event names the reason. - An ingress-nginx annotation that wxingress does not know, typos included, is ignored with a Warning event
(
... is not supported by wxingress and is ignored). - Annotations that only exist to restrict access and that wxingress does not implement make every location of the Ingress answer 503, see Not supported.
- Resources of other namespaces (Secrets, ConfigMaps) are never used, as ingress-nginx with
allow-cross-namespace-resources: false; naming one makes the locations answer 503. - Snippet annotations (
server-snippet,configuration-snippet,stream-snippet,auth-snippet) need-enable-snippets(Helmcontroller.enableSnippets), the equivalent of ingress-nginx'sallow-snippet-annotations. Without it they are ignored with a warning, and the admission webhook refuses them. wxingress has no Lua: snippets with*_by_lua*directives must be rewritten (for example in njs).
Which Ingresses are served at all is decided by their class, see -ingress-class-controllers and
-watch-ingress-without-class in Command-line flags. The
kubernetes.io/ingress.class annotation wins over spec.ingressClassName, as in ingress-nginx.
Supported¶
These behave as in ingress-nginx.
| Annotations | Notes |
|---|---|
proxy-connect-timeout, proxy-read-timeout, proxy-send-timeout |
Whole seconds, rendered with s. |
proxy-body-size, client-body-buffer-size |
client_max_body_size (k, m, g) and client_body_buffer_size (k, m). |
proxy-buffering, proxy-request-buffering |
on or off. |
proxy-buffer-size, proxy-buffers-number, proxy-busy-buffers-size |
Rendered as ingress-nginx renders them: proxy_buffer_size S and proxy_buffers N S, so either of the first two sets both (N defaults to 4, S to 4k or the ConfigMap's). |
proxy-max-temp-file-size, proxy-http-version |
proxy-http-version takes 1.0 or 1.1. |
proxy-next-upstream, proxy-next-upstream-timeout, proxy-next-upstream-tries |
ingress-nginx's validation (error, timeout, invalid_header, http_500, http_502, http_503, http_504, http_403, http_404, http_429, non_idempotent, off); non_idempotent is moved to the end. The timeout is whole seconds. |
proxy-redirect-from, proxy-redirect-to |
Validated as nginx.org/proxy-redirect-from and -to; one without the other is ignored (except off and default). |
proxy-cookie-domain, proxy-cookie-path |
proxy_cookie_domain, proxy_cookie_path: <from> <to> or off. |
rewrite-target, use-regex, x-forwarded-prefix |
A rewrite-target or use-regex turns every path of the host (of every Ingress on it) into a case-insensitive regular expression, ordered longest first, as ingress-nginx does. x-forwarded-prefix sends X-Forwarded-Prefix only from locations that rewrite-target rewrites. |
app-root, upstream-vhost, server-alias |
A server-alias wildcard must leave at least two labels (*.example.com, not *.com). |
permanent-redirect, permanent-redirect-code, temporal-redirect, temporal-redirect-code |
temporal-redirect wins. Default codes 301 and 302; a code outside 300-308 is the default, with a warning. The URL must be http or https. |
ssl-ciphers, ssl-prefer-server-ciphers |
|
backend-protocol |
HTTP, HTTPS, GRPC, GRPCS, AUTO_HTTP, FCGI (case-insensitive). Any other value (AJP) makes the locations answer 503. |
fastcgi-index, fastcgi-params-configmap |
With backend-protocol: FCGI; without it ignored with a warning. A missing ConfigMap, one of another namespace or an invalid parameter: 503. |
service-upstream |
Proxy to the Service's ClusterIP instead of its endpoints; a headless Service to its endpoints. |
enable-cors, cors-allow-origin, cors-allow-methods, cors-allow-headers, cors-allow-credentials, cors-expose-headers, cors-max-age |
ingress-nginx's defaults (origin *, methods GET, PUT, POST, DELETE, PATCH, OPTIONS, its header list, credentials true, max age 1728000) and validation; preflight requests of allowed origins are answered with 204 before auth. |
allowlist-source-range, whitelist-source-range, denylist-source-range |
IPv4 and IPv6 addresses and CIDRs. allowlist-source-range wins over whitelist-source-range. An invalid range makes the locations answer 503. |
limit-rps, limit-rpm, limit-connections, limit-burst-multiplier, limit-allowlist, limit-whitelist, limit-rate, limit-rate-after |
ingress-nginx's zones (5m, keyed by $binary_remote_addr), bursts (limit times limit-burst-multiplier, default 5, nodelay), allowlist and its 503. limit-allowlist wins over limit-whitelist. limit-rate and limit-rate-after are kilobytes (limit_rate <n>k). An invalid limit or allowlist drops the whole rate limit, as in ingress-nginx. Limits are per controller pod, as in ingress-nginx. |
auth-type, auth-secret, auth-secret-type, auth-realm |
Basic auth from an Opaque Secret (auth-file, the default, with key auth, or auth-map), updated when the Secret changes; the Authorization header is not passed on. A missing or invalid Secret, a missing auth-secret or another auth-type than basic makes the locations answer 503. |
auth-url, auth-method, auth-signin, auth-signin-redirect-param, auth-response-headers, auth-request-redirect, auth-cache-key, auth-cache-duration, auth-always-set-cookie, auth-snippet |
External auth with auth_request, as in ingress-nginx: auth-signin-redirect-param defaults to rd, auth-cache-duration to 200 202 401 5m (used with auth-cache-key). An invalid auth-url, auth-method, auth-response-headers, auth-request-redirect, auth-always-set-cookie or auth-cache-duration makes the locations answer 503. auth-snippet needs -enable-snippets. |
auth-proxy-set-headers |
A ConfigMap of the Ingress's namespace; watched. |
auth-tls-secret, auth-tls-verify-client, auth-tls-verify-depth, auth-tls-error-page, auth-tls-pass-certificate-to-upstream, auth-tls-match-cn |
Client certificates from ca.crt (and ca.crl) of the Secret: auth-tls-verify-client on (default), off, optional or optional_no_ca; depth 1 by default; auth-tls-match-cn is CN=<regular expression>. A missing Secret or an invalid auth-tls-match-cn makes the locations answer 503. |
proxy-ssl-secret, proxy-ssl-verify, proxy-ssl-verify-depth, proxy-ssl-ciphers, proxy-ssl-protocols, proxy-ssl-name, proxy-ssl-server-name |
TLS to the backends, with ingress-nginx's defaults. |
connection-proxy-header |
close or keep-alive. |
enable-access-log, enable-rewrite-log |
enable-access-log: "false" is access_log off, enable-rewrite-log: "true" is rewrite_log on. |
relative-redirects, use-port-in-redirects |
relative-redirects: "true" is absolute_redirect off; use-port-in-redirects is port_in_redirect on or off. |
satisfy |
satisfy any or satisfy all in the locations of the Ingress; another value is ignored with a warning. |
default-backend, custom-http-errors, disable-proxy-intercept-errors |
See also -default-backend-service. disable-proxy-intercept-errors: "true" stops intercepting the custom-http-errors codes; a default-backend still gets the 502 and 503. |
mirror-target, mirror-host, mirror-request-body |
mirror-request-body: off does not mirror the body (default on); mirror-host defaults to the target's host. |
ssl-passthrough |
Needs -enable-tls-passthrough (Helm controller.enableTLSPassthrough), the equivalent of ingress-nginx's --enable-ssl-passthrough; without it ignored with a warning, as in ingress-nginx. |
server-snippet, configuration-snippet, stream-snippet |
Need -enable-snippets. |
Supported with differences¶
| Annotations | Difference from ingress-nginx |
|---|---|
ssl-redirect, force-ssl-redirect, preserve-trailing-slash |
The redirect is the same (https://<host><URI>, 308 with -ingress-nginx-defaults), but whether a host redirects is decided per server, by the first Ingress of the host; ingress-nginx decides per location. |
affinity: cookie, affinity-mode, affinity-canary-behavior, session-cookie-name, session-cookie-path, session-cookie-domain, session-cookie-expires, session-cookie-max-age, session-cookie-secure, session-cookie-samesite, session-cookie-conditional-samesite-none, session-cookie-change-on-failure |
With -dynamic-upstreams: ingress-nginx's sticky balancers in the wxingress module (affinity-mode balanced or persistent, affinity-canary-behavior sticky or legacy), with the same cookie format, so existing sessions stay pinned. The cookie is INGRESSCOOKIE by default, its path the common prefix of the Ingress's paths. Without it: NGINX's sticky cookie, HttpOnly, and session-cookie-change-on-failure is always true. An nginx.org/sticky-cookie-services cookie of the same Service wins. |
load-balance, upstream-hash-by, upstream-hash-by-subset, upstream-hash-by-subset-size |
load-balance takes round_robin or ewma; upstream-hash-by-subset-size defaults to 3; with affinity both are ignored. With -dynamic-upstreams: ingress-nginx's round_robin, ewma, chash and chashsubset in the wxingress module; changing them needs no reload. Small differences: subsets are built from servers sorted by address (the same on every replica), and ewma does not average in failed tries. Without it: hash ... consistent, ewma becomes least_time last_byte, and upstream-hash-by-subset hashes to a single endpoint, with a warning. |
canary, canary-weight, canary-weight-total, canary-by-header, canary-by-header-value, canary-by-header-pattern, canary-by-cookie |
Need -dynamic-upstreams; without it, or with an invalid canary annotation, the canary Ingress is ignored with an event. canary-weight-total defaults to 100. With -ingress-nginx-defaults a canary attaches to the upstream of the matching host and path in any Ingress, as ingress-nginx's alternative backend (only in its own namespace unless -ingress-nginx-cross-namespace); without it, only to the location of the Ingress with the same host and path, not to other Ingresses that use the same Service. Weight, header and cookie changes need no reload; a canary that attaches to an upstream or leaves one reloads NGINX (ingress-nginx needs none). |
auth-type: digest |
Not available in open-source NGINX: the locations answer 503. |
auth-tls-* |
On a host without TLS: with -ingress-nginx-defaults, HTTPS with the default certificate checks the client certificate, but plain HTTP answers 503 (ingress-nginx serves it without the check). Without the defaults, the locations answer 503. An nginx.org IngressMTLS Policy wins. |
auth-keepalive, auth-keepalive-requests, auth-keepalive-timeout, auth-keepalive-share-vars |
auth-keepalive idle connections to the auth service (default 0, off), with 1000 requests and 60 s by default. As in ingress-nginx not used with HTTP/2 on (ignored with a warning) or when the host of auth-url has a variable. auth-keepalive-share-vars is what NGINX's subrequests do anyway. |
custom-headers |
A ConfigMap of the Ingress's namespace, each header listed in the ConfigMap key global-allowed-response-headers. Rendered with add_header ... always and add_header_inherit merge instead of more_set_headers, so the server's own headers stay. |
enable-cors |
Rendered with add_header ... always instead of more_set_headers. |
custom-http-errors |
Also a 502 or 503 a backend sends itself goes to the default-backend, not only NGINX's own. Without default-backend the codes go to -default-backend-service; without either the annotation is ignored with a warning. |
from-to-www-redirect |
Not applied when another resource has the other name. |
enable-opentelemetry, opentelemetry-trust-incoming-span, opentelemetry-operation-name |
With NGINX's otel module; traces go only to the ConfigMap's otel-exporter-endpoint (ingress-nginx's otlp-collector-host). Without one the annotations are ignored with a warning. |
proxy-ssl-* |
An nginx.org egressMTLS Policy wins. A Secret with neither certificate nor CA makes the locations answer 503. |
mirror-target |
The target is resolved per request with the pod's name servers. |
stream-snippet |
One file of the stream block per Ingress, removed with the Ingress. |
Not supported¶
Every ingress-nginx annotation not listed above is ignored with a Warning event. Among them are enable-global-auth
(wxingress has no global auth; ignored with a warning whatever its value, it does not make the locations answer 503), enable-opentracing and opentracing-trust-incoming-span (use OpenTelemetry), and
http2-push-preload.
Annotations that restrict access and that wxingress does not implement are not ignored: an Ingress with one of them answers 503 on all its locations, with a Warning event, because serving it without the restriction would serve what ingress-nginx refuses.
| Annotations | Answers 503 when |
|---|---|
enable-modsecurity, enable-owasp-core-rules |
set to anything but false |
modsecurity-snippet |
set |
global-rate-limit |
set to anything but 0 |
modsecurity-transaction-id, global-rate-limit-window, global-rate-limit-key, global-rate-limit-ignored-cidrs |
never on their own (they only qualify the annotations above); otherwise ignored with a warning |
Warning
Find these before switching with kubectl get ingress -A -o yaml | grep -E 'modsecurity|global-rate-limit'.
Rate limits per pod (limit-rps, limit-rpm) or a rate-limit Policy can replace a global rate limit.
Gateway API routes¶
With -enable-gateway-api, HTTPRoutes and GRPCRoutes read a subset of the same annotations, with the same code as for
Ingresses: proxy-connect-timeout, proxy-read-timeout, proxy-send-timeout, proxy-body-size,
client-body-buffer-size, proxy-buffering, proxy-buffer-size, proxy-buffers-number, proxy-busy-buffers-size,
proxy-http-version, proxy-next-upstream, proxy-next-upstream-timeout, proxy-next-upstream-tries,
service-upstream, load-balance, upstream-hash-by, upstream-hash-by-subset, upstream-hash-by-subset-size,
affinity, affinity-mode, the session-cookie-* annotations, limit-rps, limit-rpm, limit-connections,
limit-burst-multiplier, limit-allowlist (limit-whitelist), limit-rate, limit-rate-after,
canary-by-header, canary-by-header-value, canary-by-header-pattern, canary-by-cookie, backend-protocol, and
with -enable-snippets server-snippet, configuration-snippet, nginx.org/server-snippets and
nginx.org/location-snippets (without it they are ignored with a warning; an Ingress with the nginx.org ones is
rejected instead).
On a route, backend-protocol takes HTTP, HTTPS, GRPC, GRPCS and AUTO_HTTP; HTTPS and GRPCS turn on TLS
to the backends, which an egressMTLS Policy can give a client certificate and verification. The canary-by-*
annotations need -dynamic-upstreams, an HTTPRoute and a rule without a mirror: the rule's second backend is the
canary, and the backends' weights are the canary's weight. proxy-request-buffering, proxy-max-temp-file-size and
affinity-canary-behavior are accepted on a route without a warning but have no effect there.
Every other ingress-nginx annotation on a route is reported as a warning in the route's status and events and
ignored, except those that restrict access: auth-*, allowlist-source-range, whitelist-source-range,
denylist-source-range, satisfy, enable-modsecurity and enable-owasp-core-rules (other than false),
modsecurity-snippet and global-rate-limit (other than 0). As on an Ingress, a route is not served without them:
all its rules answer 503. Restrict routes with Policies attached through
wxingress.wizardtales.com/policies: <policy>[, <policy>...] on the route or the Gateway. Redirects, rewrites, header
changes and CORS are the Gateway API's own filters.
nginx.org annotations¶
The annotations of NGINX Ingress Controller keep working as upstream defines them; see the NGINX Ingress Controller annotation reference for details. Unlike the ingress-nginx annotations, most of them are validated before the Ingress is configured, and an invalid value rejects the whole Ingress. Where a ConfigMap key of the same name exists, the annotation overrides it for the Ingress, and the default is the ConfigMap's (see ConfigMap).
| Annotation | Description |
|---|---|
nginx.org/mergeable-ingress-type |
master or minion, see below. |
nginx.org/policies |
Comma-separated Policies (<name> or <namespace>/<name>) applied to the Ingress. |
nginx.org/proxy-connect-timeout, nginx.org/proxy-read-timeout, nginx.org/proxy-send-timeout |
proxy_connect_timeout, proxy_read_timeout, proxy_send_timeout (nginx time, a bare number is seconds). |
nginx.org/client-max-body-size, nginx.org/client-body-buffer-size |
client_max_body_size, client_body_buffer_size. |
nginx.org/proxy-buffering, nginx.org/proxy-buffers, nginx.org/proxy-buffer-size, nginx.org/proxy-busy-buffers-size, nginx.org/proxy-max-temp-file-size |
proxy_buffering (true/false), proxy_buffers (8 4k), proxy_buffer_size, proxy_busy_buffers_size, proxy_max_temp_file_size. |
nginx.org/proxy-hide-headers, nginx.org/proxy-pass-headers |
Comma-separated headers for proxy_hide_header and proxy_pass_header. |
nginx.org/proxy-set-headers |
Comma-separated Name: value request headers to the upstream (proxy_set_header); a name without a value sends the client's header ($http_<name>). |
nginx.org/add-header, nginx.org/add-header-inherit |
Comma-separated Name: value[: always] response headers (add_header); add_header_inherit on, off or merge. |
nginx.org/proxy-http-version |
proxy_http_version. |
nginx.org/proxy-next-upstream, nginx.org/proxy-next-upstream-timeout, nginx.org/proxy-next-upstream-tries |
proxy_next_upstream, proxy_next_upstream_timeout, proxy_next_upstream_tries. |
nginx.org/proxy-redirect-from, nginx.org/proxy-redirect-to |
proxy_redirect <from> <to>; from alone only as off or default. |
nginx.org/upstream-vhost |
The Host header to the upstream. |
nginx.org/rewrites |
serviceName=<svc> rewrite=<path>[;...]: the path the requests to that Service are rewritten to. Wins over nginx.org/rewrite-target, with a warning. |
nginx.org/rewrite-target |
The path every location of the Ingress rewrites to, with $1-style captures of nginx.org/path-regex paths. |
nginx.org/path-regex |
case_sensitive, case_insensitive or exact: the paths of the Ingress as regular expressions or exact matches. |
nginx.org/app-root |
Redirect / to this path. |
nginx.org/ssl-services, nginx.org/grpc-services, nginx.org/websocket-services |
Comma-separated Services that are proxied with TLS, with gRPC, or with WebSocket support. |
nginx.org/use-cluster-ip |
true: proxy to the ClusterIP of the Services instead of their endpoints. |
nginx.org/lb-method |
The load-balancing method, as the ConfigMap's lb-method. |
nginx.org/keepalive, nginx.org/max-fails, nginx.org/max-conns, nginx.org/fail-timeout |
keepalive of the upstreams; max_fails, max_conns and fail_timeout of their servers. |
nginx.org/upstream-zone-size |
The shared memory zone of the upstreams. No effect with -dynamic-upstreams, where all upstreams share one zone. |
nginx.org/sticky-cookie-services |
serviceName=<svc> <name> [expires=...] [domain=...] [httponly] [secure] [path=...][;...]: NGINX's sticky cookie for those Services. |
nginx.org/ssl-redirect, ingress.kubernetes.io/ssl-redirect |
Redirect plain HTTP to HTTPS on a host with TLS (default the ConfigMap's ssl-redirect). |
nginx.org/redirect-to-https |
Redirect to HTTPS based on X-Forwarded-Proto. |
nginx.org/http-redirect-code |
301, 302, 307 or 308 for the HTTPS redirect. |
nginx.org/hsts, nginx.org/hsts-max-age, nginx.org/hsts-include-subdomains, nginx.org/hsts-behind-proxy |
HSTS of the server; the other three apply only with nginx.org/hsts. |
nginx.org/http2 |
HTTP/2 on the TLS listeners of the server. |
nginx.org/ssl-ciphers, nginx.org/ssl-prefer-server-ciphers |
ssl_ciphers and ssl_prefer_server_ciphers of the server. |
nginx.org/server-tokens |
server_tokens on or off (true/false). |
nginx.org/listen-ports, nginx.org/listen-ports-ssl |
Comma-separated ports the server listens on instead of 80 and 443; not on an Ingress without a host. |
nginx.org/basic-auth-secret, nginx.org/basic-auth-realm |
Basic auth from a Secret of type nginx.org/htpasswd; the realm only with the Secret. |
nginx.org/custom-http-errors |
Comma-separated codes or 4xx/5xx of the upstreams sent to the Ingress's spec.defaultBackend (proxy_intercept_errors). |
nginx.org/limit-req-rate, nginx.org/limit-req-key, nginx.org/limit-req-zone-size, nginx.org/limit-req-burst, nginx.org/limit-req-delay, nginx.org/limit-req-no-delay, nginx.org/limit-req-dry-run, nginx.org/limit-req-log-level, nginx.org/limit-req-reject-code, nginx.org/limit-req-scale |
A request rate limit: rate <n>r/s or <n>r/m, key (default ${binary_remote_addr}), zone size (default 10m), burst, delay, nodelay, dry run, log level info, notice, warn or error (default error), reject code (default 429, 503 with -ingress-nginx-defaults), and with limit-req-scale the rate divided among the controller's replicas. |
nginx.org/server-snippets, nginx.org/location-snippets |
Lines for the server and the locations of the Ingress. Need -enable-snippets; without it the Ingress is rejected. They replace the ConfigMap's server-snippets and location-snippets for the Ingress. |
NGINX Plus and App Protect only¶
These are read only with -nginx-plus or App Protect, which wxingress does not build. An Ingress with one of them is
rejected (annotation requires NGINX Plus, annotation requires AppProtect):
| Annotation | Upstream meaning |
|---|---|
nginx.com/health-checks, nginx.com/health-checks-mandatory, nginx.com/health-checks-mandatory-queue |
Active health checks. |
nginx.com/slow-start |
slow_start of the upstream servers. |
nginx.com/jwt-key, nginx.com/jwt-realm, nginx.com/jwt-token, nginx.com/jwt-login-url |
JWT validation; use a jwt Policy, which the wxingress module validates without NGINX Plus. |
nginx.com/sticky-cookie-services |
The NGINX Plus name of nginx.org/sticky-cookie-services, which works without it. |
nginx.com/policies |
The NGINX Plus name of nginx.org/policies. |
appprotect.f5.com/app-protect-enable, appprotect.f5.com/app-protect-policy, appprotect.f5.com/app-protect-security-log-enable, appprotect.f5.com/app-protect-security-log, appprotect.f5.com/app-protect-security-log-destination, appprotectdos.f5.com/app-protect-dos-resource |
NGINX App Protect WAF and DoS. |
Where an nginx.org/* annotation and an ingress-nginx annotation set the same thing on one Ingress, the nginx.org/*
annotation wins and the ingress-nginx annotation is ignored with a warning:
| ingress-nginx annotation | Overridden by |
|---|---|
proxy-connect-timeout, proxy-read-timeout, proxy-send-timeout |
nginx.org/proxy-connect-timeout, nginx.org/proxy-read-timeout, nginx.org/proxy-send-timeout |
proxy-body-size, client-body-buffer-size |
nginx.org/client-max-body-size, nginx.org/client-body-buffer-size |
proxy-buffering, proxy-busy-buffers-size, proxy-max-temp-file-size |
nginx.org/proxy-buffering, nginx.org/proxy-busy-buffers-size, nginx.org/proxy-max-temp-file-size |
proxy-buffer-size, proxy-buffers-number |
Each by either nginx.org/proxy-buffer-size or nginx.org/proxy-buffers |
proxy-http-version, proxy-next-upstream, proxy-next-upstream-timeout, proxy-next-upstream-tries |
nginx.org/proxy-http-version, nginx.org/proxy-next-upstream, nginx.org/proxy-next-upstream-timeout, nginx.org/proxy-next-upstream-tries |
proxy-redirect-from, proxy-redirect-to |
nginx.org/proxy-redirect-from, nginx.org/proxy-redirect-to |
upstream-vhost, app-root, ssl-ciphers, ssl-prefer-server-ciphers, ssl-redirect |
nginx.org/upstream-vhost, nginx.org/app-root, nginx.org/ssl-ciphers, nginx.org/ssl-prefer-server-ciphers, nginx.org/ssl-redirect |
rewrite-target |
nginx.org/rewrite-target, nginx.org/rewrites |
use-regex |
nginx.org/path-regex |
service-upstream |
nginx.org/use-cluster-ip |
load-balance, upstream-hash-by |
nginx.org/lb-method |
affinity |
nginx.org/sticky-cookie-services, for the Services it names |
limit-rps, limit-rpm |
nginx.org/limit-req-rate |
auth-type |
nginx.org/basic-auth-secret |
server-snippet, configuration-snippet |
nginx.org/server-snippets, nginx.org/location-snippets |
backend-protocol |
nginx.org/ssl-services (for HTTPS and GRPCS), nginx.org/grpc-services (for GRPC and GRPCS) |
auth-tls-*, proxy-ssl-* |
An IngressMTLS or EgressMTLS Policy (nginx.org/policies) |
Masters and minions (nginx.org/mergeable-ingress-type) keep NGINX Ingress Controller's rules: the host belongs to
them, and other Ingresses of the host are rejected. A master ignores nginx.org/rewrites, nginx.org/ssl-services,
nginx.org/grpc-services, nginx.org/websocket-services, nginx.org/sticky-cookie-services,
nginx.org/use-cluster-ip and the health checks; a minion ignores the server-level annotations (nginx.org/hsts, nginx.org/hsts-max-age,
nginx.org/hsts-include-subdomains, nginx.org/ssl-redirect, ingress.kubernetes.io/ssl-redirect, nginx.org/redirect-to-https, nginx.org/http-redirect-code, nginx.org/server-tokens,
nginx.org/listen-ports*, nginx.org/server-snippets, nginx.org/ssl-ciphers, nginx.org/ssl-prefer-server-ciphers,
nginx.org/http2, nginx.org/app-root, nginx.org/proxy-hide-headers, nginx.org/proxy-pass-headers). Minions
inherit the master's timeouts, buffers, upstream, rate-limit, nginx.org/location-snippets,
nginx.org/upstream-vhost and nginx.org/proxy-http-version annotations unless they set their own. Regular Ingresses that share a host are merged into one server as
in ingress-nginx; see ingress-nginx compatibility.