Merge branch 'apache:master' into master

Author: hachi029

.github/workflows/build.yml

@@ -81,6 +81,10 @@ jobs:
         run: |
           bash ./ci/free_disk_space.sh
 
+      - name: Set hostname
+        run: |
+          sudo sh -c 'echo "127.0.1.1   httpbin.local" >> /etc/hosts'
+
       - name: Linux Before install
         run: sudo ./ci/${{ matrix.os_name }}_runner.sh before_install
 

.github/workflows/push-dev-image-on-commit.yml

@@ -68,7 +68,7 @@ jobs:
             "uri": "/get",
             "upstream": {
             "type": "roundrobin",
-            "nodes": { "httpbin.org:80": 1 }
+            "nodes": { "httpbin:8080": 1 }
             }
           }'
 

apisix/discovery/nacos/init.lua

@@ -22,7 +22,6 @@ local core               = require('apisix.core')
 local ipairs             = ipairs
 local pairs              = pairs
 local type               = type
-local math               = math
 local math_random        = math.random
 local ngx                = ngx
 local ngx_re             = require('ngx.re')
@@ -164,10 +163,7 @@ local function get_signed_param(group_name, service_name)
 end
 
 
-local function get_base_uri()
-    local host = local_conf.discovery.nacos.host
-    -- TODO Add health check to get healthy nodes.
-    local url = host[math_random(#host)]
+local function build_base_uri(url)
     local auth_idx = core.string.rfind_char(url, '@')
     local username, password
     if auth_idx then
@@ -195,6 +191,18 @@ local function get_base_uri()
 end
 
 
+local function get_base_uri_by_index(index)
+    local host = local_conf.discovery.nacos.host
+
+    local url = host[index]
+    if not url then
+        return nil
+    end
+
+    return build_base_uri(url)
+end
+
+
 local function de_duplication(services, namespace_id, group_name, service_name, scheme)
     for _, service in ipairs(services) do
         if service.namespace_id == namespace_id and service.group_name == group_name
@@ -277,69 +285,114 @@ local function is_grpc(scheme)
 end
 
 local curr_service_in_use = {}
-local function fetch_full_registry(premature)
-    if premature then
-        return
-    end
 
-    local base_uri, username, password = get_base_uri()
+
+local function fetch_from_host(base_uri, username, password, services)
     local token_param, err = get_token_param(base_uri, username, password)
     if err then
-        log.error('get_token_param error:', err)
-        return
+        return false, err
     end
 
-    local infos = get_nacos_services()
-    if #infos == 0 then
-        return
-    end
     local service_names = {}
-    for _, service_info in ipairs(infos) do
-        local data, err
+    local nodes_cache = {}
+    local had_success = false
+
+    for _, service_info in ipairs(services) do
         local namespace_id = service_info.namespace_id
         local group_name = service_info.group_name
         local scheme = service_info.scheme or ''
-        local namespace_param = get_namespace_param(service_info.namespace_id)
-        local group_name_param = get_group_name_param(service_info.group_name)
-        local signature_param = get_signed_param(service_info.group_name, service_info.service_name)
+        local namespace_param = get_namespace_param(namespace_id)
+        local group_name_param = get_group_name_param(group_name)
+        local signature_param = get_signed_param(group_name, service_info.service_name)
         local query_path = instance_list_path .. service_info.service_name
                            .. token_param .. namespace_param .. group_name_param
                            .. signature_param
-        data, err = get_url(base_uri, query_path)
-        if err then
-            log.error('get_url:', query_path, ' err:', err)
-            goto CONTINUE
-        end
+        local data, req_err = get_url(base_uri, query_path)
+        if req_err then
+            log.error('failed to fetch instances for service [', service_info.service_name,
+                      '] from ', base_uri, ', error: ', req_err)
+        else
+            had_success = true
+
+            local key = get_key(namespace_id, group_name, service_info.service_name)
+            service_names[key] = true
 
-        local nodes = {}
-        local key = get_key(namespace_id, group_name, service_info.service_name)
-        service_names[key] = true
-        for _, host in ipairs(data.hosts) do
-            local node = {
-                host = host.ip,
-                port = host.port,
-                weight = host.weight or default_weight,
-            }
-            -- docs: https://github.com/yidongnan/grpc-spring-boot-starter/pull/496
-            if is_grpc(scheme) and host.metadata and host.metadata.gRPC_port then
-                node.port = host.metadata.gRPC_port
+            local hosts = data.hosts
+            if type(hosts) ~= 'table' then
+                hosts = {}
             end
 
-            core.table.insert(nodes, node)
-        end
-        if #nodes > 0 then
-            local content = core.json.encode(nodes)
-            nacos_dict:set(key, content)
+            local nodes = {}
+            for _, host in ipairs(hosts) do
+                local node = {
+                    host = host.ip,
+                    port = host.port,
+                    weight = host.weight or default_weight,
+                }
+                -- docs: https://github.com/yidongnan/grpc-spring-boot-starter/pull/496
+                if is_grpc(scheme) and host.metadata and host.metadata.gRPC_port then
+                    node.port = host.metadata.gRPC_port
+                end
+
+                core.table.insert(nodes, node)
+            end
+
+            if #nodes > 0 then
+                nodes_cache[key] = nodes
+            end
         end
-        ::CONTINUE::
     end
-    -- remove services that are not in use anymore
+
+    if not had_success then
+        return false, 'all nacos services fetch failed'
+    end
+
+    for key, nodes in pairs(nodes_cache) do
+        local content = core.json.encode(nodes)
+        nacos_dict:set(key, content)
+    end
+
     for key, _ in pairs(curr_service_in_use) do
         if not service_names[key] then
             nacos_dict:delete(key)
         end
     end
+
     curr_service_in_use = service_names
+    return true
+end
+
+
+local function fetch_full_registry(premature)
+    if premature then
+        return
+    end
+
+    local infos = get_nacos_services()
+    if #infos == 0 then
+        return
+    end
+
+    local host_list = local_conf.discovery.nacos.host
+    local host_count = #host_list
+    local start = math_random(host_count)
+
+    for i = 0, host_count - 1 do
+        local idx = (start + i - 1) % host_count + 1
+        local base_uri, username, password = get_base_uri_by_index(idx)
+
+        if not base_uri then
+            log.warn('nacos host at index ', idx, ' is invalid, skip')
+        else
+            local ok, err = fetch_from_host(base_uri, username, password, infos)
+            if ok then
+                return
+            end
+            log.error('fetch_from_host: ', base_uri, ' err:', err)
+        end
+    end
+
+    log.error('failed to fetch nacos registry from all hosts')
 end
 
 

apisix/utils/log-util.lua

@@ -20,6 +20,7 @@ local expr = require("resty.expr.v1")
 local content_decode = require("apisix.utils.content-decode")
 local ngx = ngx
 local pairs = pairs
+local type = type
 local ngx_now = ngx.now
 local ngx_header = ngx.header
 local os_date = os.date
@@ -32,6 +33,7 @@ local is_http = ngx.config.subsystem == "http"
 local req_get_body_file = ngx.req.get_body_file
 local MAX_REQ_BODY      = 524288      -- 512 KiB
 local MAX_RESP_BODY     = 524288      -- 512 KiB
+local MAX_LOG_FORMAT_DEPTH = 5
 local io                = io
 
 local lru_log_format = core.lrucache.new({
@@ -69,22 +71,35 @@ local function get_request_body(max_bytes)
 end
 
 
-local function gen_log_format(format)
+local function do_gen_log_format(format, depth)
     local log_format = {}
     for k, var_name in pairs(format) do
-        if var_name:byte(1, 1) == str_byte("$") then
+        if type(var_name) == "table" then
+            if depth >= MAX_LOG_FORMAT_DEPTH then
+                core.log.warn("log_format nesting exceeds max depth ",
+                              MAX_LOG_FORMAT_DEPTH, ", truncating")
+                log_format[k] = {false, {}}
+            else
+                local nested_format = do_gen_log_format(var_name, depth + 1)
+                log_format[k] = {false, nested_format}
+            end
+        elseif type(var_name) == "string" and var_name:byte(1, 1) == str_byte("$") then
             log_format[k] = {true, var_name:sub(2)}
         else
             log_format[k] = {false, var_name}
         end
     end
+    return log_format
+end
+
+local function gen_log_format(format)
+    local log_format = do_gen_log_format(format, 1)
     core.log.info("log_format: ", core.json.delay_encode(log_format))
     return log_format
 end
 
 
-local function get_custom_format_log(ctx, format, max_req_body_bytes)
-    local log_format = lru_log_format(format or "", nil, gen_log_format, format)
+local function build_log_entry(ctx, log_format, max_req_body_bytes)
     local entry = core.table.new(0, core.table.nkeys(log_format))
     for k, var_attr in pairs(log_format) do
         if var_attr[1] then
@@ -100,10 +115,19 @@ local function get_custom_format_log(ctx, format, max_req_body_bytes)
             else
                 entry[k] = ctx.var[var_attr[2]]
             end
+        elseif type(var_attr[2]) == "table" then
+            entry[k] = build_log_entry(ctx, var_attr[2], max_req_body_bytes)
         else
             entry[k] = var_attr[2]
         end
     end
+    return entry
+end
+
+
+local function get_custom_format_log(ctx, format, max_req_body_bytes)
+    local log_format = lru_log_format(format or "", nil, gen_log_format, format)
+    local entry = build_log_entry(ctx, log_format, max_req_body_bytes)
 
     local matched_route = ctx.matched_route and ctx.matched_route.value
     if matched_route then

ci/pod/docker-compose.common.yml

@@ -111,3 +111,12 @@ services:
     restart: unless-stopped
     ports:
       - "127.0.0.1:4566:4566"            # LocalStack Gateway
+
+
+  ## httpbin - HTTP Request & Response Service
+  httpbin:
+    image: kennethreitz/httpbin
+    container_name: httpbin
+    restart: unless-stopped
+    ports:
+      - "8280:80"

docker/compose/docker-compose-master.yaml

@@ -47,6 +47,14 @@ services:
     networks:
       - apisix
 
+  httpbin:
+    image: ghcr.io/mccutchen/go-httpbin
+    restart: always
+    ports:
+      - "8280:8080/tcp"
+    networks:
+      - apisix
+
 networks:
   apisix:
     driver: bridge

docs/en/latest/plugins/clickhouse-logger.md

@@ -44,7 +44,7 @@ The `clickhouse-logger` Plugin is used to push logs to [ClickHouse](https://clic
 | timeout       | integer | False    | 3                   | [1,...]      | Time to keep the connection alive for after sending a request. |
 | name          | string  | False    | "clickhouse logger" |              | Unique identifier for the logger. If you use Prometheus to monitor APISIX metrics, the name is exported in `apisix_batch_process_entries`.                              |
 | ssl_verify    | boolean | False    | true                | [true,false] | When set to `true`, verifies SSL.                              |
-| log_format       | object  | False    |  |              | Log format declared as key value pairs in JSON format. Values only support strings. [APISIX](../apisix-variable.md) or [Nginx](http://nginx.org/en/docs/varindex.html) variables can be used by prefixing the string with `$`. |
+| log_format       | object  | False    |  |              | Log format declared as key-value pairs in JSON. Values support strings and nested objects (up to five levels deep; deeper fields are truncated). Within strings, [APISIX](../apisix-variable.md) or [NGINX](http://nginx.org/en/docs/varindex.html) variables can be referenced by prefixing with `$`. |
 | include_req_body       | boolean | False    | false          | [false, true]         | When set to `true` includes the request body in the log. If the request body is too big to be kept in the memory, it can't be logged due to Nginx's limitations.                                                                                                                                                                                 |
 | include_req_body_expr  | array   | False    |                |                       | Filter for when the `include_req_body` attribute is set to `true`. Request body is only logged when the expression set here evaluates to `true`. See [lua-resty-expr](https://github.com/api7/lua-resty-expr) for more.                                                                                                                          |
 | include_resp_body      | boolean | False    | false          | [false, true]         | When set to `true` includes the response body in the log.                                                                                                                                                                                                                                                                                        |
@@ -103,7 +103,7 @@ You can also set the format of the logs by configuring the Plugin metadata. The
 
 | Name       | Type   | Required | Default                                                                       | Description                                                                                                                                                                                                                                             |
 | ---------- | ------ | -------- | ----------------------------------------------------------------------------- | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
-| log_format | object | False |  | Log format declared as key value pairs in JSON format. Values only support strings. [APISIX](../apisix-variable.md) or [Nginx](http://nginx.org/en/docs/varindex.html) variables can be used by prefixing the string with `$`. |
+| log_format | object | False |  | Log format declared as key-value pairs in JSON. Values support strings and nested objects (up to five levels deep; deeper fields are truncated). Within strings, [APISIX](../apisix-variable.md) or [NGINX](http://nginx.org/en/docs/varindex.html) variables can be referenced by prefixing with `$`. |
 | max_pending_entries | integer | False | | Maximum number of pending entries that can be buffered in batch processor before it starts dropping them. |
 
 :::info IMPORTANT

docs/en/latest/plugins/elasticsearch-logger.md

@@ -42,7 +42,7 @@ The `elasticsearch-logger` Plugin pushes request and response logs in batches to
 | endpoint_addrs  | array[string] | True     |                             | Elasticsearch API endpoint addresses. If multiple endpoints are configured, they will be written randomly.            |
 | field         | object   | True     |                             | Elasticsearch `field` configuration.                          |
 | field.index   | string  | True     |                             | Elasticsearch [_index field](https://www.elastic.co/guide/en/elasticsearch/reference/current/mapping-index-field.html#mapping-index-field). |
-| log_format | object | False    |                             | Custom log format in key-value pairs in JSON format. Support [APISIX](../apisix-variable.md) or [NGINX variables](http://nginx.org/en/docs/varindex.html) in values. |
+| log_format | object | False    |                             | Custom log format as key-value pairs in JSON. Values support strings and nested objects (up to five levels deep; deeper fields are truncated). Within strings, [APISIX](../apisix-variable.md) or [NGINX variables](http://nginx.org/en/docs/varindex.html) can be referenced by prefixing with `$`. |
 | auth          | array   | False    |                             | Elasticsearch [authentication](https://www.elastic.co/guide/en/elasticsearch/reference/current/setting-up-authentication.html) configuration. |
 | auth.username | string  | True     |                             | Elasticsearch [authentication](https://www.elastic.co/guide/en/elasticsearch/reference/current/setting-up-authentication.html) username. |
 | auth.password | string  | True     |                             | Elasticsearch [authentication](https://www.elastic.co/guide/en/elasticsearch/reference/current/setting-up-authentication.html) password. |
@@ -61,7 +61,7 @@ This Plugin supports using batch processors to aggregate and process entries (lo
 
 | Name | Type | Required | Default | Description |
 |------|------|----------|---------|-------------|
-| log_format | object | False |  | Custom log format in key-value pairs in JSON format. Support [APISIX variables](../apisix-variable.md) and [NGINX variables](http://nginx.org/en/docs/varindex.html) in values. |
+| log_format | object | False |  | Log format declared as key-value pairs in JSON. Values support strings and nested objects (up to five levels deep; deeper fields are truncated). Within strings, [APISIX](../apisix-variable.md) or [NGINX](http://nginx.org/en/docs/varindex.html) variables can be referenced by prefixing with `$`. |
 | max_pending_entries | integer | False | | Maximum number of pending entries that can be buffered in batch processor before it starts dropping them. |
 
 ## Examples