Documentation
¶
Index ¶
- Variables
- type BatchBinding
- type BatchDB
- type Binding
- type DB
- type ElasticsearchDB
- type InfluxDB
- func (d *InfluxDB) Address() string
- func (d *InfluxDB) BatchPostfix() []byte
- func (d *InfluxDB) BatchPrefix() []byte
- func (d *InfluxDB) Concat() []byte
- func (d *InfluxDB) EntryPostfix() []byte
- func (d *InfluxDB) EntryPrefix() []byte
- func (d *InfluxDB) Init() error
- func (d *InfluxDB) Req(r io.Reader) (req *http.Request)
- func (d *InfluxDB) Translate(p []byte, w msg.Writer) error
- func (d *InfluxDB) Validate(res *http.Response) error
- type Server
Constants ¶
This section is empty.
Variables ¶
var (
// DefaultConfig is the default configuration used for NSQ communications by each binding if one is not specified.
DefaultConfig *nsq.Config
)
Functions ¶
This section is empty.
Types ¶
type BatchBinding ¶
type BatchBinding struct {
//Topic is the NSQ topic to listen on
Topic string
// Channel is the NSQ channel to listen on
Channel string
// Endpoint is the database and decode logic
Endpoint BatchDB
// Workers sets the number of concurent goroutines serving this binding; defaults to 1
Workers int
// MaxMsg sets the maximum number of messages collected before upload; defaults to 50
MaxMsg int
// BatchTime sets the maximum time spend waiting to collect messages before upload; defaults to 250ms
BatchTime time.Duration
// contains filtered or unexported fields
}
BatchBinding types connect NSQ channels, flux/msg schemas, and database endpoints, but they use database request batching.
type BatchDB ¶
type BatchDB interface {
// A BatchDB must fulfill the DB interface. The Req
// method should point the request towards the batch
// endpoint of the database, if it is different
// from the standard endpoint.
DB
// EntryPrefix is written immediately before each individual entry
EntryPrefix() []byte
// EntryPostfix is written immediately after each individual entry
EntryPostfix() []byte
// BatchPrefix is written at the beggining of the request body
// (before all entries, e.g. '[')
BatchPrefix() []byte
// BatchPostfix is written at the end of the request body
// (after all entries, e.g. ']')
BatchPostfix() []byte
// Concat is placed between every entry
// (e.g. ',')
Concat() []byte
}
BatchDB represents a database connection that handles batch uploads. XxxPrefix and XxxPostfix methods are used to format the body.
type Binding ¶
type Binding struct {
// Topic is the NSQ topic to listen on
Topic string
// Channel is the NSQ channel to listen on
Channel string
// Endpoint is the database and decode logic used for this topic & channel
Endpoint DB
// Workers sets the number of concurrent goroutines serving this binding; defaults to 1
Workers int
// contains filtered or unexported fields
}
Binding types connect NSQ channels, flux/msg schemas, and database endpoints.
type DB ¶
type DB interface {
// Init gives you an opportunity to do custom initialization and
// validation on the DB. Calls to Server.Run will cause the
// driver to call Init() exactly once.
Init() error
// Translate should turn the body of an NSQ message
// into a valid body to be used to write
// to the database.
Translate(p []byte, w msg.Writer) error
// Req should return a valid *http.Request to be performed
// by an http client. 'r' should be used
// as the body of the request.
Req(r io.Reader) *http.Request
// Validate is used to validate a response from a server
// after data is sent. Validate() should return
// a non-nil error to mark the response as failed.
// Validate is also responsible for closing the response
// body.
Validate(*http.Response) error
}
DB is the interface that fluxd uses to communicate with a database. All calls should be thread-safe.
type ElasticsearchDB ¶
type ElasticsearchDB struct {
Schema msg.Schema
Addr string
Index string
Dtype string
// contains filtered or unexported fields
}
ElasticsearchDB conforms to the DB interface. It POSTs to http://{Addr}/{Index}/{Dtype}/ using the output of Translate() as the message body.
func (*ElasticsearchDB) Address ¶
func (e *ElasticsearchDB) Address() string
Address returns the endpoint that this db POSTs to.
func (*ElasticsearchDB) Init ¶
func (e *ElasticsearchDB) Init() error
func (*ElasticsearchDB) Req ¶
func (e *ElasticsearchDB) Req(r io.Reader) *http.Request
Req returns the proper POST request to Addr/Index/Dtype
type InfluxDB ¶
type InfluxDB struct {
Schema msg.Schema
Addr string
DBname string
// contains filtered or unexported fields
}
InfluxDB implements the BatchBinding interface. It uses the first field in the Schema as the series name.
func (*InfluxDB) Address ¶
Address returns {Addr}/db/{DBname}/series?u=root&p=root, but only computes the string once.
func (*InfluxDB) Init ¶
Init must be called before the call to Server.Run() in order to initialize unexported struct members.
func (*InfluxDB) Translate ¶
Tranlsate writes InfluxDB-compatible JSON from 'p' into 'w', returning an error if it encounters a problem decoding the data. Note that the msg.PackExt type is not supported, as it cannot be written as "flat" data. The first value in the Schema is assumed to be the series name. (Any other arrangement requires a significantly more complicated implementation.)
type Server ¶
type Server struct {
// Config used for NSQ connections; defaults to DefaultConfig
NSQConfig *nsq.Config
// Address(es) of nsqlookupd(s)
Lookupdaddrs []string
// If true, requests are redirected to StdOut instead of
// to the http endpoint. Useful for testing.
UseStdout bool
// contains filtered or unexported fields
}
Server represents a collection of bindings to NSQLookupd instances.
func (*Server) BindBatch ¶
func (s *Server) BindBatch(b *BatchBinding)
BindBatch adds a batched binding to the server. Bind must be called before Run.
Source Files