more

Signed-off-by: spacewander <spacewanderlzx@gmail.com>

apisix/control/v1.lua

@@ -21,6 +21,8 @@ local get_services = require("apisix.http.service").services
 local upstream_mod = require("apisix.upstream")
 local get_upstreams = upstream_mod.upstreams
 local ipairs = ipairs
+local str_format = string.format
+local ngx_var = ngx.var
 
 
 local _M = {}
@@ -54,51 +56,103 @@ function _M.schema()
 end
 
 
-local function iter_and_add_checker(infos, values, src)
+local function extra_checker_info(value, src_type)
+    local checker = value.checker
+    local upstream = value.checker_upstream
+    local host = upstream.checks and upstream.checks.active and upstream.checks.active.host
+    local port = upstream.checks and upstream.checks.active and upstream.checks.active.port
+    local nodes = upstream.nodes
+    local healthy_nodes = core.table.new(#nodes, 0)
+    for _, node in ipairs(nodes) do
+        local ok = checker:get_target_status(node.host, port or node.port, host)
+        if ok then
+            core.table.insert(healthy_nodes, node)
+        end
+    end
+
+    local conf = value.value
+    return {
+        name = upstream_mod.get_healthchecker_name(value),
+        src_id = conf.id,
+        src_type = src_type,
+        nodes = nodes,
+        healthy_nodes = healthy_nodes,
+    }
+end
+
+
+local function iter_and_add_healthcheck_info(infos, values, src_type)
     if not values then
         return
     end
 
     for _, value in core.config_util.iterate_values(values) do
         if value.checker then
-            local checker = value.checker
-            local upstream = value.checker_upstream
-            local host = upstream.checks and upstream.checks.active and upstream.checks.active.host
-            local port = upstream.checks and upstream.checks.active and upstream.checks.active.port
-            local nodes = upstream.nodes
-            local health_nodes = core.table.new(#nodes, 0)
-            for _, node in ipairs(nodes) do
-                local ok = checker:get_target_status(node.host, port or node.port, host)
-                if ok then
-                    core.table.insert(health_nodes, node)
-                end
-            end
-
-            local conf = value.value
-            core.table.insert(infos, {
-                name = upstream_mod.get_healthchecker_name(value),
-                src_id = conf.id,
-                src_type = src,
-                nodes = nodes,
-                health_nodes = health_nodes,
-            })
+            core.table.insert(infos, extra_checker_info(value, src_type))
         end
     end
 end
 
 
-function _M.healthcheck()
+function _M.get_health_checkers()
     local infos = {}
     local routes = get_routes()
-    iter_and_add_checker(infos, routes, "routes")
+    iter_and_add_healthcheck_info(infos, routes, "routes")
     local services = get_services()
-    iter_and_add_checker(infos, services, "services")
+    iter_and_add_healthcheck_info(infos, services, "services")
     local upstreams = get_upstreams()
-    iter_and_add_checker(infos, upstreams, "upstreams")
+    iter_and_add_healthcheck_info(infos, upstreams, "upstreams")
     return 200, infos
 end
 
 
+local function iter_and_find_healthcheck_info(values, src_type, src_id)
+    if not values then
+        return nil, str_format("%s[%s] not found", src_type, src_id)
+    end
+
+    for _, value in core.config_util.iterate_values(values) do
+        if value.value.id == src_id then
+            if not value.checker then
+                return nil, str_format("no checker for %s[%s]", src_type, src_id)
+            end
+
+            return extra_checker_info(value, src_type)
+        end
+    end
+
+    return nil, str_format("%s[%s] not found", src_type, src_id)
+end
+
+
+function _M.get_health_checker()
+    local uri_segs = core.utils.split_uri(ngx_var.uri)
+    core.log.info("healthcheck uri: ", core.json.delay_encode(uri_segs))
+
+    local src_type, src_id = uri_segs[4], uri_segs[5]
+    if not src_id then
+        return 404, {error_msg = str_format("missing src id for src type %s", src_type)}
+    end
+
+    local values
+    if src_type == "routes" then
+        values = get_routes()
+    elseif src_type == "services" then
+        values = get_services()
+    elseif src_type == "upstreams" then
+        values = get_upstreams()
+    else
+        return 400, {error_msg = str_format("invalid src type %s", src_type)}
+    end
+
+    local info, err = iter_and_find_healthcheck_info(values, src_type, src_id)
+    if not info then
+        return 404, {error_msg = err}
+    end
+    return 200, info
+end
+
+
 return {
     -- /v1/schema
     {
@@ -110,6 +164,12 @@ return {
     {
         methods = {"GET"},
         uris = {"/healthcheck"},
-        handler = _M.healthcheck,
+        handler = _M.get_health_checkers,
+    },
+    -- /v1/healthcheck/{src_type}/{src_id}
+    {
+        methods = {"GET"},
+        uris = {"/healthcheck/*"},
+        handler = _M.get_health_checker,
     }
 }

doc/control-api.md

@@ -89,7 +89,7 @@ Return current [health check](health-check.md) status in the format below:
 ```json
 [
     {
-        "health_nodes": [
+        "healthy_nodes": [
             {
                 "host": "127.0.0.1",
                 "port": 1980,
@@ -113,7 +113,7 @@ Return current [health check](health-check.md) status in the format below:
         "src_type": "upstreams"
     },
     {
-        "health_nodes": [
+        "healthy_nodes": [
             {
                 "host": "127.0.0.1",
                 "port": 1980,
@@ -146,4 +146,35 @@ Each entry contains fields below:
 object with id 1 creates a health checker, the `src_type` is `upstreams` and the `src_id` is `1`.
 * name: the name of the health checker.
 * nodes: the target nodes of the health checker.
-* health_nodes: the health node known by the health checker.
+* healthy_nodes: the healthy node known by the health checker.
+
+User can also use `/v1/healthcheck/$src_type/$src_id` can get the status of a health checker.
+
+For example, `GET /v1/healthcheck/upstreams/1` returns:
+
+```json
+{
+    "healthy_nodes": [
+        {
+            "host": "127.0.0.1",
+            "port": 1980,
+            "weight": 1
+        }
+    ],
+    "name": "upstream#/upstreams/1",
+    "nodes": [
+        {
+            "host": "127.0.0.1",
+            "port": 1980,
+            "weight": 1
+        },
+        {
+            "host": "127.0.0.2",
+            "port": 1988,
+            "weight": 1
+        }
+    ],
+    "src_id": "1",
+    "src_type": "upstreams"
+}
+```

doc/health-check.md

@@ -22,6 +22,9 @@
 Health Check of APISIX is based on [lua-resty-healthcheck](https://github.com/Kong/lua-resty-healthcheck),
 you can use it for upstream.
 
+Note that we only start the health check when the upstream is hit by a request.
+There won't be any health check if an upstream is configured but isn't in used.
+
 The following is an example of health check:
 
 ```shell
@@ -105,3 +108,5 @@ contains: `active` or `passive`.
   * `passive.unhealthy.tcp_failures`: Number of TCP failures in proxied traffic to consider a target unhealthy, as observed by passive health checks.
   * `passive.unhealthy.timeouts`: Number of timeouts in proxied traffic to consider a target unhealthy, as observed by passive health checks.
   * `passive.unhealthy.http_failures`: Number of HTTP failures in proxied traffic (as defined by `passive.unhealthy.http_statuses`) to consider a target unhealthy, as observed by passive health checks.
+
+The health check status can be fetched via `GET /v1/healthcheck` in [control API](./control-api.md).

doc/zh-cn/health-check.md

@@ -23,6 +23,9 @@
 
 APISIX的健康检查使用[lua-resty-healthcheck](https://github.com/Kong/lua-resty-healthcheck)实现，你可以在upstream中使用它。
 
+注意只有在 upstream 被请求时才会开始健康检查。
+如果一个 upstream 被配置但没有被请求，那么就不会有健康检查。
+
 下面是一个检查检查的例子：
 
 ```shell
@@ -105,3 +108,5 @@ curl http://127.0.0.1:9080/apisix/admin/routes/1 -H 'X-API-KEY: edd1c9f034335f13
     * `passive.unhealthy.tcp_failures`: 如果TCP通讯失败次数超过 `tcp_failures` 次，则将upstream节点设置为 `unhealthy` 状态。
     * `passive.unhealthy.timeouts`: 如果被动健康检查超时次数超过 `timeouts` 次，则将upstream节点设置为 `unhealthy` 状态。
     * `passive.unhealthy.http_failures`: 如果被动健康检查的HTTP请求失败（由 `passive.unhealthy.http_statuses` 定义）的次数超过 `http_failures`次，则将upstream节点设置为 `unhealthy` 状态。
+
+健康检查信息可以通过 [控制接口](./control_api.md) 中的 `GET /v1/healthcheck` 接口得到。

t/control/healthcheck.t

@@ -87,6 +87,14 @@ upstreams:
                 return a.host < b.host
             end)
             ngx.say(json.encode(res))
+
+            local code, body, res = t.test('/v1/healthcheck/upstreams/1',
+                ngx.HTTP_GET)
+            res = json.decode(res)
+            table.sort(res.nodes, function(a, b)
+                return a.host < b.host
+            end)
+            ngx.say(json.encode(res))
         }
     }
 --- grep_error_log eval
@@ -95,7 +103,8 @@ qr/unhealthy TCP increment \(.+\) for '[^']+'/
 unhealthy TCP increment (1/2) for '(127.0.0.2:1988)'
 unhealthy TCP increment (2/2) for '(127.0.0.2:1988)'
 --- response_body
-[{"health_nodes":[{"host":"127.0.0.1","port":1980,"weight":1}],"name":"upstream#/upstreams/1","nodes":[{"host":"127.0.0.1","port":1980,"weight":1},{"host":"127.0.0.2","port":1988,"weight":1}],"src_id":"1","src_type":"upstreams"}]
+[{"healthy_nodes":[{"host":"127.0.0.1","port":1980,"weight":1}],"name":"upstream#/upstreams/1","nodes":[{"host":"127.0.0.1","port":1980,"weight":1},{"host":"127.0.0.2","port":1988,"weight":1}],"src_id":"1","src_type":"upstreams"}]
+{"healthy_nodes":[{"host":"127.0.0.1","port":1980,"weight":1}],"name":"upstream#/upstreams/1","nodes":[{"host":"127.0.0.1","port":1980,"weight":1},{"host":"127.0.0.2","port":1988,"weight":1}],"src_id":"1","src_type":"upstreams"}
 
 
 
@@ -141,6 +150,14 @@ routes:
                 return a.port < b.port
             end)
             ngx.say(json.encode(res))
+
+            local code, body, res = t.test('/v1/healthcheck/routes/1',
+                ngx.HTTP_GET)
+            res = json.decode(res)
+            table.sort(res.nodes, function(a, b)
+                return a.port < b.port
+            end)
+            ngx.say(json.encode(res))
         }
     }
 --- grep_error_log eval
@@ -149,7 +166,8 @@ qr/unhealthy TCP increment \(.+\) for '[^']+'/
 unhealthy TCP increment (1/2) for '127.0.0.1(127.0.0.1:1988)'
 unhealthy TCP increment (2/2) for '127.0.0.1(127.0.0.1:1988)'
 --- response_body
-[{"health_nodes":[{"host":"127.0.0.1","port":1980,"weight":1}],"name":"upstream#/routes/1","nodes":[{"host":"127.0.0.1","port":1980,"weight":1},{"host":"127.0.0.1","port":1988,"weight":1}],"src_id":"1","src_type":"routes"}]
+[{"healthy_nodes":[{"host":"127.0.0.1","port":1980,"weight":1}],"name":"upstream#/routes/1","nodes":[{"host":"127.0.0.1","port":1980,"weight":1},{"host":"127.0.0.1","port":1988,"weight":1}],"src_id":"1","src_type":"routes"}]
+{"healthy_nodes":[{"host":"127.0.0.1","port":1980,"weight":1}],"name":"upstream#/routes/1","nodes":[{"host":"127.0.0.1","port":1980,"weight":1},{"host":"127.0.0.1","port":1988,"weight":1}],"src_id":"1","src_type":"routes"}
 
 
 
@@ -200,6 +218,14 @@ services:
                 return a.port < b.port
             end)
             ngx.say(json.encode(res))
+
+            local code, body, res = t.test('/v1/healthcheck/services/1',
+                ngx.HTTP_GET)
+            res = json.decode(res)
+            table.sort(res.nodes, function(a, b)
+                return a.port < b.port
+            end)
+            ngx.say(json.encode(res))
         }
     }
 --- grep_error_log eval
@@ -208,7 +234,8 @@ qr/unhealthy TCP increment \(.+\) for '[^']+'/
 unhealthy TCP increment (1/2) for '127.0.0.1(127.0.0.1:1988)'
 unhealthy TCP increment (2/2) for '127.0.0.1(127.0.0.1:1988)'
 --- response_body
-[{"health_nodes":{},"name":"upstream#/services/1","nodes":[{"host":"127.0.0.1","port":1980,"weight":1},{"host":"127.0.0.1","port":1988,"weight":1}],"src_id":"1","src_type":"services"}]
+[{"healthy_nodes":{},"name":"upstream#/services/1","nodes":[{"host":"127.0.0.1","port":1980,"weight":1},{"host":"127.0.0.1","port":1988,"weight":1}],"src_id":"1","src_type":"services"}]
+{"healthy_nodes":{},"name":"upstream#/services/1","nodes":[{"host":"127.0.0.1","port":1980,"weight":1},{"host":"127.0.0.1","port":1988,"weight":1}],"src_id":"1","src_type":"services"}
 
 
 
@@ -224,3 +251,21 @@ unhealthy TCP increment (2/2) for '127.0.0.1(127.0.0.1:1988)'
     }
 --- response_body
 {}
+
+
+
+=== TEST 5: no checker
+--- request
+GET /v1/healthcheck/routes/1
+--- error_code: 404
+--- response_body
+{"error_msg":"routes[1] not found"}
+
+
+
+=== TEST 6: invalid src type
+--- request
+GET /v1/healthcheck/route/1
+--- error_code: 400
+--- response_body
+{"error_msg":"invalid src type route"}